LaTeX vs Word; improvements of LaTeX over the years

When comparing MS Word vs LaTeX the proponents of Word often say that many advantages of LaTeX over Word from say 1997 disappeared over the years as Word evolved to the current 2013 version. For example,

• Word no longer has a binary format, the files are just zipped XML files;
• In Word you can easily define templates and structure your text using defined levels of headings, etc.;
• Word 2013 is much more stable for large documents than it was 10 years ago;
• Word now has an improved equation editor with nearly plain-text Unicode editing and customizable keyboard shortcuts.

Now, LaTeX too has improved over the years. So my question is the following:

What are major improvements of LaTeX or associated tools (e.g., editors) which

• fix some (major) disadvantages LaTeX might have had in comparison to Word some years ago?
• or introduce new (killer-) features where Word can't keep up?

I want to restrict the question to changes from 2000 to today. Please add a short description of the improvement/feature and the year it was introduced.

• Compared to an earlier time, more LaTeX style packages exist for out-of-the-box solutions to common typesetting challenges (problems or formats). Dec 18, 2014 at 15:46
• I have used Word since version 1.1 and regarding templates, the best version was version 2.0c. It was notorious unstable (crashed at least ten times a day), but it did not unexpectedly change my template. From version 2002, I gave up templates because they always changed in ways I could not understand. Neither could my IT department, but the reason was some weird interaction between word, the network and how they set up mobile computing. Word is best for a single user on a single computer, for a advanced user in a restricted network environment,... impossible. Typesetting nice document? Not .. Dec 18, 2014 at 16:09
• Is it easy, in Word, to write equations? How about inline math? Is it easy to write a document template? Do you still have to struggle against the software when trying to create a bulleted or numbered list? Word may indeed have shown substantial improvement, but this is not the same as it being worthwhile to use. If the best thing to be said about Word documents is that you don't have to use Word to edit them... why wouldn't one use LaTeX instead? Dec 18, 2014 at 16:11
• "Nowadays Word doesn't have a binary format anymore, the files are just zipped xml files" – Well, the specification for those "just zipped XML files" weighs in at 6546 pages, and has to this day not been fully implemented, not even by MS Office itself. Dec 18, 2014 at 20:04
• @Formagella Word equations are incredibly clunky to write. There's no excuse for having to use a mouse to build equations (or edit them). Dec 19, 2014 at 1:51

I do not know of LaTeX ever having any disandvantages as compared to Word over the years, other than Word had more funding, so I will only comment on the new (killer) features where Word can't keep up.

1. Macros - been there from day one, include here hundreds of new packages.

2. PGF/TikZ

3. XeTeX and LuaTeX, ConTeXt

4. Microtype

5. The ability to exercise your brain while you're typing and writing macros

6. The confidence that your code and documents can survive possibly for ever.

7. A great community.

• You can't think of a single disadvantage of LaTeX over Word? How about the fact that, for simple documents, anyone can just open up Word and create something adequate. LaTeX requires you to learn stuff before you can even write "Hello world." Dec 19, 2014 at 14:02
• @DavidRicherby "anyone" can get a pen and write on a piece of paper. There is no causation to disantvantages or advantages in your statement. You need to look at what you want to achieve. If your aim is to typeset a good looking book then the "ease part" does not play a role. Dec 19, 2014 at 15:17
• The causation is that Word has a user interface that makes it easy and intuitive to do simple things in simple documents, whereas LaTeX has a user interface that requires learning multiple different steps to produce even a simple document. Yes, that learning pays for itself many times over but it still has to be done. It appears that the reason you see no advantages at all in Word is that you've restricted your attention only to use cases (e.g., typesetting books) where LaTeX is a clear winner. Dec 19, 2014 at 15:30
• A steep learning curve cannot be seen as a good thing unconditionally, because there's simply no limit (why not write pdf documents in an hex editor?). Maybe the good outweighs the bad for latex, and on this website, most will probably agree on that; and maybe the good of latex requires some learning curve at some point (but I am not convinced of that, see for example lyx or texmacs). In any case, it doesn't mean that the learning curve, by itself, a good thing. Dec 19, 2014 at 16:47
• @DavidRicherby: Define 'anyone' and 'something adequate'. Most people, maybe, but I can't. Now I'll grant that I'm several sigmas away from the normal/average user, but my every attempt at using Word, or more frequently Open/LibreOffice, to do more than read a document someone else has sent me has been a dismal failure. Their learning curves are, for me, more like cliffs, while LaTeX is just another programming language. Dec 19, 2014 at 19:23

this is not about any improvements in (la)tex. it is only about stability.

it is very likely that an article created (and published) in 2000 in latex can be re-used directly today in a volume of "collected works", whereas such an article created at the same time in word will most likely require considerably more work, possibly even rekeying.

mathematics has a long shelf life.

• I really like the shelf life part. Dec 18, 2014 at 16:45
• Hm... might work today with a plain document from 2000, but I am not so sure about it in 2030 with a LaTeX document from today. I have experienced too many problems reusing documents less than three years old if they employed modern packages or classes, such as biblatex, libertine, tikz, moderncv. Unfortunately, issues with back-compatibility are no longer a privilege of Word and Co. Dec 18, 2014 at 20:48
• @Daniel -- true enough. however, if you keep track of what versions of various packages were used (the ams uses the snapshot package to record that information), and keep an archive of old versions, reprocessing old versions is possible with little hassle. it's too bad that many package writers don't try to enforce backward compatibility. it's essential in a production environment. the minimum requirement is good documentation of incompatible changes. Dec 18, 2014 at 21:28
• @CarlWitthoft: Just for curiosity I did: No problems at all to open it in a current version of Word. Dec 19, 2014 at 8:10
• I have twice had papers that I was preparing in Word fall into some unrecoverable state from which the only way to avoid retyping the whole thing was to cut and paste bits and pieces of the text into a new document. In one case I couldn't even cut and past directly; it had to go through Notepad first to clean out all the formatting. So the shelf life of the Word document wasn't even long enough to publish it once. This hasn't happened in about five years, but that's mainly because I convert my Word drafts to LaTeX before anything could go wrong with them. Dec 27, 2014 at 3:14

The main reason I gave up Word for LaTeX is so that I could keep all of my documents under version control. You could argue that I could do the same thing with a .docx file, but Git likes .tex just fine, and I can do all the usual cool Git tricks as well.

• This is a highly underrated point! XML or no, the amount of cruft inside a .docx (or .doc, or .docm) document makes version control near-impossible. Dec 19, 2014 at 1:57
• Doesn't office provide built-in version control? Dec 19, 2014 at 11:37
• @.T. Verron - If it ain't Git, it ain't version control. (No offense, personal opinion) Dec 19, 2014 at 14:14
• @T.Verron how good is Word at merging what three authors did, all starting from version A? How much work is conflict resolution? Git does that better than, say, svn or perforce (in my opinion) - and if Word played in the same ballpark, I’d be really impressed. Dec 20, 2014 at 10:40
• @T.Verron: If it's built-in into Office, can I still integrate non-Office content that I want to manage together with my documents in the same version control? How about version-controlled projects that, for one reason or another, do not include any Office documents at all - can I still use the built-in Office version control for those? Dec 20, 2014 at 20:06

The main advantage I see with LaTeX is that everything's in the open. You can ask a question, like on StackExchange, by quoting straightforward text, and you get an answer in straightforward text rather than "try clicking this menu and that, oh, in Windows 7 this may be in a different ribbon and they self-organize anyway". Word help tends to work better with somebody in the same room.

So as social networks and wikis and sites like StackExchange have grown in importance over the last few years, the advantage of LaTeX being something that one can competently talk about and view and discuss in plain text media has grown.

• I think this is a very good point! +1 Dec 18, 2014 at 20:27
• So true. Try googling your *Office problems... Dec 18, 2014 at 23:59
• @Raphael I have - there is often very high quality documentation available from Microsoft. Word's GUI makes figuring out unfamiliar tasks much easier as well. Dec 19, 2014 at 14:24
• Word has a "what do you want to do?" input field which lets you use a command even if you do not know where it is exactly. Agreed, it is still more complicated than a mere copy and paste in your preamble. Dec 19, 2014 at 17:10
• @Superbest: Maybe the tasks you have been trying to find were not "unfamiliar" enough. I have frequently failed to find any suitable information about something I was trying to do in Word ... and when I do find a description, half of the time it turns out it describes the menu structure of the pre-ribbon versions, and everything is named and sorted differently now. Dec 20, 2014 at 20:04

Let's break down the list first;

• Nowadays Word doesn't have a binary format anymore, the files are just zipped xml files

This sounds as if it is a good thing.

I wish. The gist is there but if you really want to use it on long documents, better buy lots of antidepressants or start a boxing course. A simple problem example that my company almost everyday suffers; We use a commercial version system and different people update the same document. Almost always the table of contents on the side bar is repeated if there are hyperlinks in it. If you search online, you get wishful thinking solutions from MVP MCSE YMCA bla bla certified experts. So far we don't know why, we just accepted it. Many say there is a page break somewhere but even if you delete everything it's not fixed. Probably, main template XML gets corrupted somewhere. So yes it is there but I won't touch it.

• In Word you can easily define templates and structure your text using for defined levels of headings etc.

This is indeed the case but without the easily part. It really got improved over the years. But again, the <insert very strong language here> preceeding pagebreaks are breaking not only the page but the document if you are 0.tiny careless.

• Word 2013 is much more stable for large documents as it was 10 years before

A dung beetle can roll stuff that is 10 times its own weight but still it's a tiny shit that it is carrying. When you really scale up, this is certainly not true. Double click a huge document, go wash your car, come back... still counting words, spell checking bla bla... fill in your tax form, post it, come back... maybe it is ready. Now insert a page break or restructure the document, wait for the government to respond to the tax form, come back... Maybe.

• Improved equation editor with nearly plain-text unicode-editing and customizable keyboard shortcuts

I'm not going to take this one seriously. See the dung beetle argument for the improvement part. The \theta is still almost 45 degrees slanted. In fact all math renders as if there is a wind blowing from the left edge affecting only the math. Even our software guys complain when they read our specifications.

But! The field management and auto-completing the templates are really improving. Finally I can feel like something is automated in Word after so many years. Sure there are some quirks but still it is really getting better. Also font management is somewhat better. It doesn't look as if your screen has been shaven with old razors anymore. And to be honest every once in a while it is almost there with the kerning. But definitely still not there. Another thing that should be mentioned is the cross-referencing. It's not fun to refer to things but at least it discovers referable material much better. Captions are still just a joke I think.

Finally, my list of the recent improvements of TeX and da Dudes,

• L3 project kicked off (I really really appreciate all the efforts)
• Normal people :) realized that TeX is actually a (horrible :P ) language and it exploded the interest and demand which in turn gave us a lot of fancy packages.
• Beamer is being (ab)used more and more. I'm watching a lot of software-cons lately for learning how-to-code and seeing lots of beamer presentations.
• fontspec is a life saver. Lua- and Xe- TeX saved a lot of time and effort.

Actually here is an example, if some MS Word person bothers you about TeX, give this problem to them to fix in word. How to suppress the operation of a luatex-defined macro on a string if the string is part of macro or a label and the final progress New package, selnolig, that automates suppression of typographic ligatures. That is simply amazing to me. just like hyperref and biblatex. There is no way you can match them via Word.

• Imo the first modern answer in this thread. :) Just one thought: while making changes to large documents with latex is easy, but compiling them can take some time. Word is slow to start, but then you have fast (or maybe not, based on other answers here) feedback on your changes, while with latex you can have instant start (depending on your editor), but then the feedback loop is slow. Dec 19, 2014 at 17:22
• @T.Verron Big documents are usually structured so you can work on it part by part via \input{} or \include{} otherwise we would see a pike in the PhD suicides :) So far what I'm really fed up with is Ctrl+A + F9 + click about ten times entire document and after finishing a subitem you hit enter twice then you change your mind and want to create it again but fun begins. Oh my, I can really punch my computer every time I have to do that. Dec 19, 2014 at 17:42
• Many thanks for mentioning the selnolig package. :-)
– Mico
Dec 19, 2014 at 20:43

High level decisions

Two major improvements of the last ten years:

1. The large packages (KOMAscript, memoir, pstricks, TiKz, e.g.) matured. The major improvement wasn't the definition of hundreds of new commands or subpackages. The major improvement was the development of high level commands which made writing in LaTeX less »programming while writing«. Users without technical background now can realise documents according to their own taste, without the need of ugly hacks.

2. MikTeX and TeXlive are easy to install, well maintained systems, really user friendly.

You could mention hundreds of really valuable developments, take the whole Oberdieck bundle as an example. Thanks to Heiko the colourstack in pdftex no longer paints large parts of a page by accident! And yes, there is LuaTeX, the LaTeX3 project and so much more.

After ten years of writing with LaTeX, the advantage I really appreciate is that my documents have more text and less LaTeX, although they have a more sophisticated layout.

• Thanks for giving the first answer which I think actually answers the users question! Dec 22, 2014 at 23:53

Word has two big ‘advantages’ over LaTeX: The ability to (a) let you smash together terrible looking documents in minutes, and (b) totally distract you from the content of the document (the text) when you try to correct the mess.

Well, this was not about the improvements of LaTeX.

The biggest improvement in LaTeX the last years, is the much better documentation that is available for free on the internet. Also, internet sites like StackExchange and similar, which make it possible to ask questions and get immediate response to problem, are big improvements, making it possible for mortals to utilize one of the best feature of LaTeX:

The ability to write solid and consistent templates that will not change without you implementing that change.

Of course, I may list lots of disadvantages of LaTeX, but that was not the question.

• +1 for the mention of TeX.SX - its existence is a huge improvement. Before I found this site I didn't know anyone with much more (La)TeX knowledge than me. That isn't the case any more... Dec 18, 2014 at 19:45
• (b) is not longer true: Word 2013 (maybe also 2010) supports different "views" of the document: a view for printing, a view for web publishing, etc. This includes a view for focussing on the contents, with minimal emphasis on the text layout? (a) is a matter of taste of course, but a lot of effort has been put in improving what people will get if they don't care, and word documents created with a recent version of word, plus two clicks to choose a theme, will maybe not be to your taste, but not be that terrible either. Dec 19, 2014 at 11:36
• @T.Verron The outline view goes back at least to MS Word 6 (1993 vintage). I don't know how useful it was for large documents, because I did not produce large documents back then.
– user
Dec 19, 2014 at 12:03
• @MichaelKjörling I'm not very familiar with word, so I wasn't saying when it got introduced. What is in 2013, and wasn't in 2003, is that, as far as I can see, this approach is somewhat encouraged when you create a new document. Dec 19, 2014 at 12:56
• @Sveinung i.imgur.com/0H1YlJm.png This is 15-20 minutes with word, and most of the time has been spent trying to guess the interface (I'm definitely not an experienced word user). Making it an exact copy of the booktab example would be a matter of finding the relevant parameters in the manual and applying them (the long part being the first). For example, applying the 1.5pt rules on top and bottom took me a few clicks (select the parameters, click the top rule, click the bottom rule), which is faster than what I'd have needed in latex (finding the top of the table, entering "top"...) Dec 22, 2014 at 12:14

Most of the greatest recent advantages of LaTeX have been mentioned, but I want to highlight the fact that XeTeX and LuaTeX are among the best systems for multi-lingual documents. In my line of work, I have found that the only truly reliable way of composing documents in different languages, scripts and even dialects is with recent TeX engines and the associated packages. Indeed, even the classic pdfTeX is now excellent in handling multiple languages and scripts — and that is no small feat considering that it is an ‘8-bit’ engine without native Unicode support.

I once read that even Adobe InDesign does not natively support complex typesetting of many eastern languages in their original scripts, e.g. Arabic, and that one must separately purchase an extension to InDesign in order to get everything that is needed (I don’t know if this is still true, though). I won’t even get started on complex typography in Word.

In modern TeX Live, one faces a beautiful dilemma:

• with pdfTeX, the very best typesetting with all microtypographic features and classic PostScript fonts, plus unparalleled stability, or
• with XeTeX and LuaTeX, easier integration of XML sources, native support for Unicode and OpenType and excellent font management packages.

One final rhetorical question: does Word 2013 finally support all OpenType features?

• Answer to rhetorical question: no, but neither does any TeX system. Dec 20, 2014 at 7:09
• What are these "all OpenType features"? Dec 20, 2014 at 17:48
• Is that true ? A chinese-japanese-vietnamese-french document with multilingual section titles is not really easy to achieve in my experience. I really dislike word, but I must admit that using characters from vietnamese, chinese and japanese scripts is much easier.
– TZDZ
Dec 22, 2014 at 15:41
• Indeed, not all OpenType features, but a reasonably complete subset. One of my biggest gripes with Word is that, in the 2010 version, it still had no support for true small caps, and I don’t know for sure if they are available in Word 2013, but apparently no. As for documents with the CJK languages, vietnamese, and the like, I have never tried, but I would hope there is a way to make it work with Fontspec and Polyglossia? Dec 24, 2014 at 17:41

We are comparing the incomparable. Word is a user interface for creating documents and TeX is a language for typesetting. We can say that Word is something like Lyx, but it means that the most important aspect is the user interface. Word generates a special XML language. But the TeX language is much more flexible than this XML. And very important is that the TeX language was designed for direct interaction with human beings when writing a document (specially when good markup for human is devised for example with good plain TeX macros). On the other hand, the XML needs the inter-layer between human and the generated XML in all circumstances, because this is its design. This mentioned inter-layer is Word.

The main advantage of TeX was, is and will be: we can avoid this inter-layer and write/modify/search etc. the TeX code directly. And the TeX language is very flexible (LaTeX is only one example of this language flexibility).

Of course, I don't mention about open vs. commercial software because this was mentioned by other answers. I never in my life created any document in Word. All my documents are only in TeX. This is a slight complication for me because I am working in lawmaking in Academic senate and Council of Higher Education Institutions in the Czech republics. Everything which I get from these activities is in Word and the first thing I do is convert those to something more reasonable. I did do a generator based on TeX for the publishing of all Czech laws, for example.

• They have pretty comparable end products. Dec 19, 2014 at 7:48
• @percusse The end products depend on the quality of used fonts and used style for typographical design. This is most visual feature of end product. On the other hand you are right: the paragraph breaking algorithm, microtypography etc. influence the end product too. Dec 19, 2014 at 8:09
• I'd say @percusse's point was mainly that the two programs have the same scope; hence they are comparable somehow. Their approach is, as you said, rather different though. Dec 19, 2014 at 12:25

I'm very leery of (collaborative) open-source software which is supposedly better than much more polished commercial software full of "bloat and eyecandy". For a long time I avoided LaTeX, since I assumed the users were just being elitist about an overcomplicated system (needless to say, I no longer hold this opinion). Even so, I have recently (after the release of Word 2013!) started using LaTeX where possible. This was explicitly due to killer features LaTeX has, that Word not only doesn't have, but likely cannot possibly have.

I still use Word for quick, short, simple documents. The extra overhead for writing and maintaining documents in LaTeX unfortunately makes itself felt. But I've learned the hard way that there is a threshold (usually around 10 pages for me) after which it's best to give up on Word and switch over to LaTeX.

Yes, but only insofar as it allows you to claim that it does in this exact sort of discussion, and compels me to begrudgingly accept that yes, technically it does.

However, these features are extremely basic. If you're a caveman who has just heard of ToCs and want to include one in your new great Pangean novel, it will work. But realistically, many texts are written to strict requirements. There are many nuances of formatting, style and just the particulars of how these are generated that Word is unable to do. There are many more that Word is able to do, but it is so laborious as to be impractical.

Point and click is nice for using features for the first time. But once you label your 1000th figure, it becomes very annoying that you can't just press some simple key combination to quickly get a label that looks exactly like you want (unless you happen to want the default one).

In Word you can easily define templates and structure your text using defined levels of headings, etc.

This has, in fact, been possible at least as far back as Word 2003. The problem with styles in WYSIWYG is that there is no longer a 1-to-1 map between appearance and style, so you can't always easily tell exactly which characters are using which style. For instance, is this period italic (it's not irrelevant - if you start inserting characters after an italic period, Word will make them italic)? Is this space bold? Is this bit of italic text actually using the Emphasis style, or is it just italic (Word sometimes can match the styles automatically, but this is not reliable)?

On the other hand, when editing markup there is in fact a 1-to-1 correspondence between what you see on the screen and the actual structure of the text - because every formatting command will be appropriately indicated in the source (which is what you will be working with anyway). The trade-off is that you can no longer easily see the appearance of the output easily (exception: features like the buggy in-line preview in TexStudio), but it turns out that for many documents, it's not hard to "imagine" how a given bit of TeX looks after you've been working with that document for an extensive time.

Also, a word on defaults: The default styles of Word tend to be inadequate. To my untrained eyes, they are only mediocre. My trained graphic designer friends say they are hideous. The extra styles you could look for on the internet and download are, for the most part, also not very pretty. Conversely, even a default document in TeX looks gorgeous. Sure, you could easily make a "LaTeX style" in Word 2013 - but as a user who hadn't seen how nice LaTeX looks and isn't trained I had assumed that professional typesetting can only be done by large publishing companies, and never even tried with Word.

Word 2013 is much more stable for large documents than it was 10 years ago

Indeed, in my experience it hardly ever crashes. In that sense it is stable.

Performance is a whole other matter. With TexStudio, I see hardly any impact of length on performance. Even compile times can be kept manageable thanks to include. With Word, after adding a couple dozen figures and tables, a few hundred references, typing into the middle of the text leads to very perceptible and distracting lag.

Word now has an improved equation editor with nearly plain-text Unicode editing and customizable keyboard shortcuts.

The equation editor in Word is unfortunately crap. Again, typesetting a very basic equation for the sake of being able to say that yes, Word can technically typeset equations is possible and easy.

But once you start trying to do real work, you quickly notice gaps in functionality, both in terms of what forms can be displayed and whether the tools provided are effective at saving you time and effort. Add to this the issue I mentioned above about seeing exactly what the input markup was from looking at the output, and it becomes challenging and frustrating to get your equations looking just right. Sometimes (like a simple sheet of notes from a non-math talk) it doesn't matter if the equations look perfect. But other times (like writing a paper) it may be very important or even mandatory.

As for keyboard shortcuts - well named TeX commands are mnemonics of themselves, and if your editor supports auto-completion, things become much easier. Keyboard shortcuts have to be laboriously memorized and are less straightforward.

What are major improvements of LaTeX or associated tools (e.g., editors) which fix some (major) disadvantages LaTeX might have had in comparison to Word some years ago? or introduce new (killer-) features where Word can't keep up?

I think development of LaTeX editors like TexStudio has made working with LaTeX much easier for many users, as well as drastically lowering the barrier to entry. Sites like this one have made getting help much easier.

One killer feature that was a big factor for me, which you don't mention, is working with parts of the document at a time. If you have a long book, and want to work on one chapter at a given time, Word does not make this easy. Keeping the whole thing in one file causes lag and reduces the usefulness of the scrollbar. You could have separate documents, but then how will the page numbering of Book - Chapter 6 know to stay in sync with Book - Chapter 5? How will you keep the styles consistent? If you edit your document a lot, and accurate formatting is critical (e.g. grades taken off for formatting errors), just validating the formatting of your document becomes a headache in it's own right.

A second one is figures. The one Word feature I miss in TeX is being able to drag and drop JPEGs to exactly where I want them - but aside from that, Word is very annoying in how it jiggles your figures around the text when you add content in preceding sections, or how there is not a tight coupling between figure and caption (which is still just an ordinary textbox, that the caption tool positions for you initially).

Another small one is citation styles. Word has a few citation styles included by default, but there are glaring omissions (for example, the style used by Nature). You can luckily download some common ones if you search online, but if you still cannot find the one you need, making your own citation style is not a simple matter.

Speaking of citations, the built in reference manager in Word is horrendous for anything beyond a handful of cites. If you have more than 5 sources, the busywork of filling out a whole form for every source really adds up, meanwhile tools like JabRef can just fetch all the information by parsing output from sites like Google Scholar. I'm not sure if this counts as a missing feature, since it is solved by those same TeX tools - for instance, JabRef can export citations for Word, effectively nullifying this problem.

• Regarding drag and drop: For emacs it seems to be possible in principle (but it doesn't work for me): tex.stackexchange.com/questions/44126/drag-and-drop-in-auctex, see also tex.stackexchange.com/questions/59575/… and emacs.stackexchange.com/questions/3955/… Dec 19, 2014 at 16:55
• @student It turns out you can do it in TexStudio as well - but at this point I have gotten used to my own \newcommands and it is actually less convenient to drag and drop. Dec 20, 2014 at 0:59
• thank you for mentioning the math input. i guess i'm not thinking of the equation editor per se, but the revamped mechanism introduced after the addition of thousands of math symbols and alphanumerics to unicode. people complain about the number of braces required by {la}tex to keep things straight. (i complain about the number of braces, preferring delimited macros.) but word uses parentheses for both structuring and nesting, and it's not easy to know whether a paren is just a structural marker or will appear in the output. yikes! Dec 20, 2014 at 20:27

I don't think this qualifies as an improvement, since it's always been that way, but for me the big advantage of LaTeX over just about anything is that I can write documents using the same editor I use for code, and so don't have to retrain my fingers to an entirely different set of commands.

For actual improvements, I'd say the ability to go straight to pdf format, and to use that to get near-immediate feedback on what the actual document looks like, by running e.g. 'gv --watch' on the document.

• Another huge advantage that has always been that way is that what you delete is really deleted, not just hidden from less sophisticated recipients of your document. Dec 18, 2014 at 18:51
• +1 I love SyncTeX, which lets you hop from the code to the right place in the PDF and back. Dec 19, 2014 at 15:34
• Actually I'd say that the one in your fist paragraph is a big disadvantage of Emacs and VIM: their keyboard shortcuts haven't standardized to the same set as the rest of the software ecosystem (ctrl+C, ctrl+V), so you have to retrain your fingers. Dec 24, 2014 at 10:00
• @Federico Poloni: That's only true if the " rest of the software ecosystem" you use uses those particular commands. For me, it isn't: at a rough guess, at least 95% of my software-using time is spent in an ecosystem - xterm, editor, gdb, etc - that doesn't use those "standardized" commands. You know what they say about standards: everyone should have their own :-) And in fact, offhand the only place I really see that "standard" is in Word-like programs (OO, LibreOffice, etc) and web browsers. Dec 25, 2014 at 7:05
• @FedericoPoloni: what is the "rest of the software ecosystem"? My "software ecosystem" consists mainly of Emacs and a bootloader for it (I'm using that small piece of software called "Ubuntu" for that purpose). Dec 29, 2014 at 1:34

• Your grandmother can probably use it. (You may sneer, but not every human is a computer expert. Some of them are salesman or piano tuners.)

• It's highly compatible. (I.e., almost everybody has it — especially in a commercial setting.)

• You can instantly see what the final document is going to look like. You don't have to "compile" it first.

• It's easy to quickly throw together trivial documents and make them look approximately how you want.

• It has a spell-checker! (And a grammar checker that erroneously tells you to reword everything in the active voice.)

• Lots of programs that aren't Word know how to import or export it.

• It's trivial to paste an Excel worksheet or chart into a document, or indeed to insert lots of other external content.

• Final output looks ugly. (Newer versions have improved this, but the defaults still look pretty ugly.)

• Doesn't handle large documents well.

• Virtually impossible to separate content from style, or to change styles without manually applying them to thousands of pages one at a time by hand.

• Cross-referencing never works properly (unless you remember to constantly click "update cells" all the time).

• Numbering lists should be trivial, but isn't.

• Equations are just painful.

• Documents sometimes get "corrupted" in ways where editing suddenly does very strange things (e.g., delete one letter, and half the page becomes bright green, or changes font, or two paragraphs further up suddenly vanish...). Sometimes a document becomes so corrupted that opening it crashes Word. (Usually opening with LibreOffice instantly fixes this — which is amusing...)

• You can't put it in version control. (Well, you can, but despite what somebody said about it being "just text" now... no, it's basically a binary file. Any edits transform the underlying XML beyond recognition, so you can't meaningfully diff it.)

• With default settings, the output is stunningly beautiful. MS Word has never looked even remotely this pretty.

• Excellent support for equations.

• Solid support for cross-referencing.

• You can trivially version-control it.

• Multiple people can work on different parts of a TeX document simultaneously. (Because you can split it into multiple source files.)

• If you have text-mangling tools (e.g., cpp), you can easily generate multiple "versions" of the same document from a single source. (Closest thing Word has is mail merge.)

• You can (to a limited extent) define your own semantic constructs, styles, etc., and change them instantly be altering their definitions.

• You have to be a computer programmer to use it. (Seriously, your grandmother wouldn't have a clue.)

• TeX doesn't understand graphics, and it doesn't understand colour. You can have these things, but don't expect them to work reliably.

• TeX also doesn't understand hyperlinks, but that tends not to be so much of an issue. (It's supposed to produce printed documents, remember?)

• If you find a mistake in the printed output, it might take ten minutes to work out where that is in the source code.

• There is no spell-checker. (And even if you could find a 3rd party spell-checker, it would just complain endlessly that \newcommand isn't a word.)

• Syntax errors can make the entire document unusable. (Word doesn't have this issue — although, document corruption issues...)

• Memorising things like \varepsilon is a lot harder than just "Insert Symbol..." and scroll until you see it. (Then again, TeX has way more symbols, so...) On the plus side, once memorised it is faster to type. (Also, need to remember which package(s) to import to get that symbol!)

• It's drastically harder to rearrange the contents of a document. Like, if you decide you want that table on the left side instead, oh and put that other paragraph at the top of the page instead... in a WYSIWYG environment, that's a simple case of drag and drop, mostly. With TeX it's a complicated cut and paste job. Heaven forbid you want to reorder the columns of a table!

• The default styling is excellent, but it's almost impossible to change anything. If you want your HTML document in Helvetics instead of Times Roman, it's a trivial one-line CSS change. If you want to change any detail of what TeX does... good luck with that!

• Occasionally TeX does something you don't want it to, and nothing in heaven and earth will make it stop. (E.g., when it insists on putting a page break where you don't want one.)

• Packages sometimes fight with each other, sometimes with dire consequences.

For me, the biggest problem with TeX is its clunky, long obsolete macro-expansion programming language, where styling is defined imperatively rather than declaratively. I have long wished we could take the excellent formatting engine and font system from TeX and bolt it into something with CSS styling, support for graphics and colours, and sane syntax. But I don't see that happening any day soon...

In short, I use TeX because the output is so excellent. The input is pretty awful, though.

At my last job, we had two employees whose job was to copy / paste-special thousands of chromatagrams into giant 14,000-page Word documents. As far as I know, there's simply no way for a non-technical user to do that kind of thing with TeX. You would have to screen-shot the spectrometer software, save the image files, convert them to the correct format for import into TeX, and then write the TeX code to generate the rest of the report... Much as I wanted us to start using TeX instead of Word, never going to happen.

When it comes down to it, MS Word is designed primarily for editing documents. TeX is only concerned with typesetting them. It's for producing the final product. So, to some extent, this is all an apples-to-oranges comparison.

• Hi and welcome, this is quite a good answer. I am disagreeing a bit on disadvantages of (La)TeX. Since LaTeX is a macro language, it is easy for one guy to define macros another human beeing with the ability to read can use. There are spellcheckers out there with a special TeX-mode (if they don't recognize the file ending). Dec 19, 2014 at 20:37
• I think you have not been using new TeX packages in the last 5 years. Practically your first 5 points are not true about disadvantages of TeX. Plus who writes down those documents to be handled by an editing documents software in the first place? Dec 19, 2014 at 20:45
• CSS is a markup language, just like LaTeX You are using styles someone else designed. If that isn't enough, define own styles. In my POV, there is not really a difference. Dec 19, 2014 at 21:01
• There is package graphics to support graphics, package graphicx to support a better interface for users. Package color and xcolor for support of colors (with a nice interface). Package hyperref takes care of, well, hyperlinks. I think this is what @percusse meant. Dec 19, 2014 at 21:02
• @MathematicalOrchid That thing about color is not true. There is simply ample amount of packages which fixes that. Better not hit on that nail if you mean LaTeX. You can have as many pages as you want. xcolor for example is a top notch package. Dec 20, 2014 at 0:11

...all the 9 answer before me +

We can program with/in TeX

1. Comand line: It is easy to include in programs, scripts, makefiles

2. (Any → Tex): tex is textual: we can write tools (DSL in Perl, Python, Ruby, yacc, etc) that generates LaTeX or parts of LaTeX.

3. (Tex → Any): tex is textual: we can write tools that convert to other type of documents (TeX4ht, pandoc, etc)

4. (include (Any → Pdf | Tex)): You can write TeX-styles that use external commands to process in LaTeX commands (eg abc, gnuplot, graphviz, ...)

5. It is a very powerful tool for PDF transformation (pdfpages, graphicx, etc): a good way to reorder pages, transform, cut, build proceedings from individual articles

and it has a good sty for chess typesetting :)

and it has a transcendental version number :)

A comment some comments in this "question" are too good to be just comments and should be reformated as answers!!

Edit 1

About 3: can *TeX be exported to docx? ;)

yes: pandoc x.tex -o x.docx

The bad news: it works for hello-word LaTeX. But I would not try "the LaTeX companion" :)

• You do now have enough reputation to comment.
– user
Dec 19, 2014 at 12:17
• About 3: can *TeX be exported to docx? ;) Dec 19, 2014 at 17:24
• @Verron: Please see edit 1. :) Dec 19, 2014 at 17:54
• Parsing TeX for converting it to something else is, and has always been, a mess. Sure, if all you have is the occasional \emph{} it works, but otherwise it's nearly impossible. I love the idea behind TeX, but I'd really appreciate a newer typesetting language with a less quirky syntax. Dec 24, 2014 at 10:03
• Actually, the version number is not yet transcendental, but a sequence of rational numbers approaching a transcendental number with each new version. Dec 26, 2014 at 10:28

There is one huge advantage of Latex over word or anything GUI type program like it:

For generating large documents, I learned how to write a program which generates the Latex document itself. This comes very handy, when processing thousands of equations and results and making tables and doing analysis on the data.

I write the program which generates the Latex in Maple or Mathematica, which does the computation, and analysis, and as it does computation, it generates Latex code along the way into one file. At the end, I end up with one huge index.tex file, ready to be compiled to pdf and html (thanks to tex4ht) with makefile.

This is the huge advantage of automation. Scripting, programs, all working together to produce a document which would take me weeks or months to type by hand, but now I do in few short hours by running a program which generates the document itself.

Try this with word or indesign. One will be sitting pointing and clicking all day long.

Nice thing, if I want to modify something, I just modify the program that generates the Latex and run it again. I just generated a pdf that contains over 700 results from computation last night, with Latex output for everything. Over 300 pages. Took 3 hrs to produce. Before that I wrote program which generated over 700 pages file for 2000 differential equations analysis. If I had to use word to type this, I will still be on the second page clicking with the mouse and doing manual edits on the screen.

edit: to answer comment for examples: Here are 2 examples. All these were generated by a program which generated the Latex. example 1 and example 2 everything there was written in Latex. Even the web pages (my web site is all Latex, thanks to tex4ht). I compile it all in makefiles on Linux. Try to build a web site using Word :)

I thought to add a diagram of one such program.

• Do you have a published example of such a large project? Dec 19, 2014 at 18:29
• Holy mother of pearls that's a crazy load of formulas, tables and calculations! I have to agree, you would probably be dead by now if you tried to do that by hand in Word. Dec 19, 2014 at 19:17
• small oddity on the web page of example 1 -- the "ff" ligatures (in the html text) are italic. (probably wouldn't have noticed it if the word "differential" weren't so prominent.) Dec 19, 2014 at 20:30
• @barbarabeeton the "ff" is know issues. Please see why-ff-displays-strange-using-unicode-encoding-vs-iso-8859-1-in-html-output-f Dec 20, 2014 at 0:21
• @Ruslan -- i didn't have any trouble accessing the two examples, either the linked page or the pdf pages linked from there. so you must have some sort of restriction specific to your site. Dec 21, 2014 at 13:13

Tl;dr: A few new features have brought latex closer to word, but word retains a few advantages. Latex still has unmatched killer major features, but none of them is really new.

fix some (major) disadvantages LaTeX might have had in comparison to Word some years ago?

Answers should start with these disadvantages, maybe. Here are a few I can think of (or got proposed by Word users to me):

• Support for system fonts The user doesn't want to know if his font is opentype, or truetype, or type3, he just knows he installed it on his system and expects to be able to use it. This works out of the box in every word version I ever used. This does not work with latex.

This is no longer a shortcoming of *tex: xetex and lualatex have the package fontspec to fill that gap.

• Learning curve Online documentation (including this site) has improved a lot, but we still are not at the point where a newbie can fire-up his editor and come with a basic document without effort.

• Click and point for complicated layout Can be seen as either a pro or a con... Yet, layout tends to break with minor edits in word, but the same is true in LaTeX: a cautious user, taking care of the content before polishing the layout will have no problem in either system.

As an extreme example (and out of word), making a presentation with beamer and tikz is funny and all, but it is much more time consuming than patching three blocks in powerpoint.

In the same way, typesetting a table in word is much easier than it is with LaTeX. Sure, maybe you won't get the nice features you want (and with recent versions of office, you actually probably will get them), but is the occasional detail worth the constant overhead?

• Support for in-document annotation In LaTeX, if someone wants to annotate your document, he either needs to do it on the pdf (and you will have to alternate between the pdf and the source) or on the source (but then he needs to know latex). In word, this problem doesn't exist (provided the reviewer has word). Essentially, this boils down to the "learning curve" problem.

or introduce new (killer-) features where Word can't keep up?

Same here, what would be the killer features of LaTeX?

• Text-readable files Not new, but true. Note that it is only true for your documents, not those you find on the internet: usually, you have only access to the pdf file, and this one is as binary as docx.

• Edittable binary files can get corrupted The problem stands for pdf as well (at least theoretically... I have never heard of a pdf file getting corrupt. Maybe that's because edittable facilities of pdf are much more restricted). So again a win for *tex.

• Availability of the software Word now has an stripped down online version, which is both free (as in beer) and cross-platform. But you need internet access. Incidentally, I find that projects like writelatex, making users forget about words like "compilation", offer some form of convergence between word and latex on the internet battlefield.

• Never spend hours trying to get the pagebreaks correct Sure. But have you really never spent hours trying to debug some document that simply won't compile? Or removed your packages, one by one, because some online platform uses its own compiler and doesn't have an up-to-date texlive? So that's a draw here, imo.

• Semantic formatting, bibliography, references, math typesetting I'm told word has made a lot of progress here, but that's not personal experience.

• If you haven't ever seen a corrupted PDF, then you haven't seen much :) chat.stackexchange.com/transcript/41?m=18977588#18977588
– yo'
Dec 20, 2014 at 0:34
• Hm, I didn't mean an unopenable pdf, that can happen for a variety of reasons. But a pdf file that one day works well, and then, after a harmless change (on a pdf, this is rather restricted, but it can be an annotation, for example), gets corrupted, that I have never seen. And yes I haven't seen much. Dec 20, 2014 at 6:03
• Well, I'm not sure if you'd succeed in annotating a corrupted PDF. But still, they exist, so I suppose it's possible that it's corrupted in such a way that a program is willing to add the annotation, but it makes the corruption as bad as in my country's government and the file can't be opened anymore.
– yo'
Dec 20, 2014 at 8:02
• Usually, corrupted means that it can't be opened, yes. But have you ever seen a (truly) valid pdf file be opened, annotated and saved, and become corrupted (=unopenable) in the process? Dec 20, 2014 at 8:06
• to me, "corrupted" doesn't mean just "can't be opened". if the content of the file becomes unreadable -- which can happen if fonts aren't embedded, and i've seen that lots of times! -- it's pretty useless anyhow, even if someone can add annotations. Dec 20, 2014 at 13:13

This is not really a LaTeX/TeX improvement, but probably existed because TeX existed: the literate programming term coined by Donald Knuth.

Although the concept is not so new (the Literate Programming book was published in 1992), this allowed in the last years an unprecedented relation between stored data, analysis, programming and manuscripts, IMHO mainly because of the collaboration of R and LaTeX, first due to the Sweave function of R and lately with knitr.

In words of Friedrich Leisch, author of Sweave;

The purpose of Sweave is to create dynamic reports, which can be updated automatically if data or analysis change. Instead of inserting a prefabricated graph or table into the report, the master document contains the R code necessary to obtain it. When run through R, all data analysis output (tables, graphs, . . . ) is created on the fly and inserted into a final LaTeX document. The report can be automatically updated if data or analysis change, which allows for truly reproducible research .

Now reproducible research is a emerging totem in sciences, probably because the basic paradigm of "write the report around the computation analysis" makes this science requirement really available in practice.

For Word, the alternative usually is still "digest" the result of a hard job with the statistical program, handwrite in Word some untraceable means and p-values, copy & paste some graphs, or even remake manually the graph in Excel to receive a comfortable feeling ("Why this "odd" boxplot? A bar graph will look better, sure ...") And when data changes, the best to do is send the manuscript to the trash and start again.

Really, the literate programming concept has surpassed the limits of the TeX universe. One can also make dynamic reports in Word in many ways, since because of the increasing popularity of R, you can obtain statistical reports with other formats such as markdown, HTML, ODT or RTF that can exported or used by Word directly, as shown in the Task of Reproducible Research in CRAN. But this is not the usual way to do things by the average Word user (In my limited experience).

Moreover, on one hand, the other formats could be useful for other purposes (such as showing results in a Web page or not having to learn LaTeX) but not to produce high quality ready-to-publish dynamic articles. In this sense, there are not alternatives as far I know. On the other hand, literate programming with LaTeX is not limited to Sweave and R. With the noweb tools you can integrate any programming language in LaTeX. I do not have experience except with R, so edits at this respect (or any other) are welcomed.

• great observation, if not, strictly speaking, a direct answer to the question. (hope the edits have been a positive addition.) Dec 20, 2014 at 13:42
• @barbarabeeton Very positive. (N.B: I must re-read what I write often).
– Fran
Dec 20, 2014 at 15:42

Long-term interoperability is the first in honor for (La)TeX.

Word cannot boast in this, for example, W'07 docs show no warranty to be open totally properly by W'10.

• I sometimes find latex file which use packages I don't have, or are require different versions to the one I have installed, and I century opened a Word 2 (written on windows 3.1!) Document fine in the latest version of word. Dec 22, 2014 at 23:57
• Long-term interoperability hasn't improved since 2000 has it? Dec 28, 2014 at 12:46
• @bubba Well, there is new evidence for the longer-termness of it!
– cfr
Dec 29, 2014 at 2:02
• @ChrisJefferson My Word 2.0 documents written in a Norwegian version of Word 2.0 does not open perfectly in either version of Word after 2.0. For example, all date fields are broken. Aug 2, 2015 at 21:33

From a point of view of a user who uses Latex only as a tool while being mostly ignorant of the underlying processes I've noticed two significant changes:

The time to resolve a problem has significantly reduced from 10 years ago when I started. The main reason is that there is much more to the point information on the net. Furthermore, what nearly disappeared where the 'read the manual' replies to people who had the same question as I. (In such cases I usually either read the manual and didn't get forward or the manual has about 100 pages.) Today, I often find those simple questions answered concisely. In short, the attitude and level of support of the community has changed considerably to the better.

I have the impression that the tools I use for my documents work together much better. The workflow is immensely smoother. gnuplot (cairolatex), asymptote, pdflatex (KOMA/revtex, hyperref, utf8, fontenc), biblatex (biber) just work without having to care or tinker at all. This also includes the publishers side, where automatic scripts seem to do a great job getting the work into the intended form.

Both aspects increased perceived productivity considerably and reduced frustration.

• Programmability. You can use VBA to change a Word document (e.g. to change its layout), but it's a pain in the ass. There are so many inconsistencies, limitations, commands that should work but don't etc. Writing code for Word is an exercise in frustration.
• Consistency. It's difficult to enforce a consistent layout in a Word document. You can define styles and templates, but users are free to deviate from the defined styles. Autonumbering is a convoluted mess and prone to failure. Advanced layout requires the use of macros (see above).
• Typesetting quality. Word limits what you can do with styles. Keeping paragraphs with the next paragraph is possible, but only barely. It's a crude mechanism compared to what's possible in other tools. Exact positioning isn't possible.
• *TeX users are also free to "deviate from the defined rules" and achieve arbitrarily ugly text. One of the best examples being the style sheets some journal designed to match their word format. Dec 19, 2014 at 11:39
• AIU, the template designer can deviate from the rules, but the users can only choose from the styles created by the template designer. Dec 19, 2014 at 11:51
• Both with word and with latex, users can do whatever they want: {\color{red}\huge\textbf{1. THE SECTION TITLE}} Dec 19, 2014 at 12:59
• @Hobbes -- users can and do deviate. i was once asked by the editor of an ams journal (he also happened to be the ams president at the time) why some articles were so ugly. when i vetted the input, every deviation from "expected" was a "correction" because the author thought he knew better. (the deviations were all in the math and they were really ugly!) one writes and distributes the best, clearest guidelines one can, but many authors are pleased to say that they never read instructions. Dec 19, 2014 at 13:10
• -1: none of these is a post-2000 improvement. Dec 24, 2014 at 10:07

I'm not sure that all of these occurred after 2000, but here are a few improvements that are fairly recent (which is what the OP asked about).

(1) Easy authoring tools like TeXworks and TeXShop. Type some stuff, click on the "Typeset" command, and you get a PDF file. No messing around with DVI or with command-line jibberish.

(2) The font mess seems to have been fixed. Things like fontspec and unicodemath are vast improvements, in my opinion.

(3) The memoir package has finally allowed me to create documents that match my tastes, rather than matching the tastes of Donald Knuth or Leslie Lamport.

(4) OpenType math fonts. This is not a TeX improvement (in fact, it owes more to Microsoft than to the TeX community), but it has the potential to significantly improve math typesetting, and reduce the amount of manual fiddling required to get things to look the way you want.

(5) A single source of authoritative help -- this site. It used to be the case that there were umpteen different places where you might (or might not) get your TeX questions answered. This site now seems to be the dominant one, and I've almost always been able to get the help I needed.

Though the OP didn't ask, I think that (as always) there are still things that need to be improved. A few examples are:

(1) Package management. The answer to every question seems to be "there's a package ...", but these packages seem to conflict in mysterious ways.

(2) Real-time typesetting. Personally, I don't like the two-step type-then-compile workflow. I want to see what the document will look like as I'm typing it. The idea that TeX will take care of all appearance issues for me is naive.

(3) Tables. Still much easier in Word.

(4) The religious fervour of the community. Many TeX/LaTeX users seem to think that all other solutions (especially commercial ones) are universally inferior and inherently evil. This attitude makes it difficult to have intelligent discussions, sometimes.

• regarding memoir, the first edition of the manual was released 3 june 2001, so yes, post-2000. re package management, almost everything has been done by volunteers; only a few projects (except for ctan and tex live) are funded or produced by formal organizations, and their focus is not primarily tex. yes, coordination could be better, but there isn't enough time or interest among the volunteers for such a massive undertaking without some sort of support. Dec 28, 2014 at 13:35
• regarding real-time typesetting, are you willing to have your ability to type input slowed down when your project exceeds some (large) number of pages? computers are faster than they used to be, but there is still a penalty for imaging every character as it is entered, especially when reflowing paragraphs and pages is involved. no, tex will not take care of all appearance issues, but a final, careful reading -- on paper -- is still the best way to decide how the product needs to be refined. Dec 28, 2014 at 13:41
• @barbarabeeton: regarding package management -- the cost of the project doesn't reduce its value, it only reduces the ROI. I'd be willing to pay, but I suspect many people would not. It's ironic that one of TeX's most highly praised attributes (it's free) places a strong constraint on its development. Dec 29, 2014 at 0:38
• @barbarabeeton: I don't accept that real-time typesetting as I type would lead to unbearable slowness. The whole document does not need to be typeset, only the part I'm looking at. I agree that final reading on paper is a good way to do a final check, but people who are much younger than you or me might disagree -- lots of people never print anything, these days. Dec 29, 2014 at 0:44
• @bubba: Regarding real-time typesetting: It would distract me a lot if something would compile constantly in the background - I want to compile when I'm ready to check my code. But if you really need it try bakoma tex (it is not free, and imho it doesn't use one of the new engines xetex/luatex.) Dec 29, 2014 at 10:20

I will try to give a clear answer as many answers here point out interesting things but I don't find them clear enough or not answering the actual question.

What are major improvements of LaTeX or associated tools which fix some disadvantages LaTeX might have had in comparison to Word some years ago?

• Speed: Compilation got much faster thanks to faster CPUs, larger RAM and SSDs. Therefore compiling a document takes fraction of what it used to take. From this other improvements follow:

• Better editors: Editors like TeXstudio (2009), Texmaker (2003) enable you to write LaTeX in the left pane and show the compiled pdf in the right pane. So you get instant feedback and thanks to synctex (2008) the position in the pdf and the source are synchronised! So the feedback you get in LaTeX is kind of comparable to Word.
• Online LaTeX editors: Servers (and the Internet) got faster and one can do much more in a browser nowadays. Therefore it is possible to create, share and collaboratively edit LaTeX online (see ShareLaTeX (2013) for instance). This is where LaTeX is on the line with Google Docs or Microsoft Office Online, but still maintains its power. Great tool if a couple of people from various scientific institutes need to write a paper together.
• Other tools: Some tools such as PDFjam (2002) that use LaTeX internally emerged due to speed improvements.
• Tools: As mentioned already above, the tools for LaTeX manipulation got much better:

• They provide feedback on the look of your document.
• Debugging information is much more helpful, usually you are pointed to the line with error.
• WYSIWIG wizards for common case scenarios have been created (e.g. table wizard in TeXstuio).
• Menus/buttons for common tasks such as font size, font style, common environments... Even standard shortcuts work like Ctrl+B, Ctrl+I.
• Character and symbol catalogs are built-in the tools so you don't have to remember (or search for) codes of symbols you don't use often.
• New tools emerged, such as Freescobaldi (2009) - LaTeX music editor, LyX (1995) - WYSIWYM LaTeX editor, Wikipedia uses LaTeX to render equations, there is even a Facebook chat plugin which enables LaTeX equations in the chat.
• User friendliness:

• Manuals and documentation got much better. No more man or info commands, but nice, clear pdfs with documentation, online tutorials or TeX@StackExchange.
• Installing LaTeX is quite easy with TeXLive (1996) or MiKTeX (??). Also, MikTeX has automatic package management, so new LaTeX packages will be installed automatically when needed.
• LaTeX has also became a standard in the scientific community, so it is now taught on most of the universities and hence general knowledge of LaTeX is growing.
• TeX/LaTeX improvements

• Tons of new packages, literally for everything. Although drawing easy diagrams or changing fonts is a piece of cake in Word, packages lower Word's dominance in these areas slightly.
• Export to LaTeX is more and more supported by other programs such as Inkscape (2003) or GeoGebra (2001) which offer export to TikZ. So you get the comfort of WYSIWIG and then export it to LaTeX where you have 100% control.
• New compilers such as XeTeX (2004) of LuaTeX (2007) provide support UTF-8 characters, custom font support, etc. XeTeX can use any fonts installed in the OS, which makes it even with Word. Honestly, this is a very important and vital improvement for a lot of users.

What are major improvements of LaTeX or associated tools which introduce new (killer-) features where Word can't keep up?

I think LaTeX has its fundamental features from the start (i.e. plain-text format, flexibility, correct typography, strict separation of input and output, price, etc.) and nowadays it is mostly trying to improve on the features it lacks compared to Word or other tools. LaTeX and Word will happily coexist, as each of them is a tool designed for a different thing. LaTeX might get some nice editor that will be comparable to Word and that will help it spread. LaTeX might get support for faster (parallel?) compilation. LaTeX might even transform into something totally with simpler and more modern syntax. We will see.

The question was about recent improvements to LaTeX (since the year 2000), but most of the answers are just comparisons of LaTeX capabilities with MS Word, regardless of when those capabilities were implemented. Since several such answers are highly up-voted, this seems to be what people want, so I'll add my own two cents worth here.

It seems to me that the pro/con arguments on this topic are often based on dubious reasoning. Some people just hate anything from Microsoft, or they think that all software should be free, so they have reasons to prefer LaTeX regardless of its capabilities. Other advice is based on folk-lore that seems to get repeated over and over again, even though it's false (in my opinion). The folk-lore includes things like:

(1) Word can only represent the appearance of a document, whereas LaTeX represents its logical structure. This is false. You can represent document structure using Word "styles", too. LaTeX forces you to think about structure; Word allows this, and even encourages it, but does not demand it.

(2) LaTeX gives you "separation of format and content" -- you can just type text, and you don't have to worry about formatting and appearance. This is not really true. LaTeX does a lot of formatting automatically, but it can't do everything. Even the books by Donald Knuth are full of little "tweaks" that he used to improve appearance.

(3) MS Word math looks horrible. This is somewhat a matter of taste and convention, so people will have to decide for themselves whether Word math looks nice. Word's math layout algorithms are derived from the TeX ones, and Word can even do some fine adjustments that TeX can't because it makes more use of the capabilities of OpenType math fonts. Anyway, this is a matter of taste/choice.

(4) MS Word math is slow because it requires too many mouse clicks. In fact, recent MS Word equation editors (MathType, or the built-in one) all allow you to type Tex-ish codes if you want to (and can remember them). Simple in-line math might be a bit faster in LaTeX, but not much. For complex formulae, I find LaTeX slower because, when I make mistakes, the only output I get is a list of mysterious error messages that take time to decipher. So, even when I'm writing a LaTeX document, I often use a graphical equation editor (MathType) that outputs LaTex code. If I have to go look up the LaTeX code for a symbol, then LaTeX is obviously a lot slower. Even if I know the code, clicking on an arrow icon is faster than typing \leftrightarrow (for me, anyway).

While the four claims listed above are dubious (in my view), LaTeX does have some real advantages. And, in a community like this, there's no shortage of people to point them out, so there's no need for me to repeat them. My purpose is to provide a bit of balance.

It's interesting to ask the Word-vs-LaTeX question in different places. The answers you get are quite different -- every community has its own bias and prejudices, obviously. In fact, in some places, the most common answer is "what's LaTeX?".

Disclaimer: I consider myself to be an expert user of MS Word. I've been using it since 1987, and I've written thousands of pages of stuff, including documents that are several hundred pages long, containing lots of mathematics. With LaTeX, I'm at the beginner/intermediate level. I have been writing a book using LateX for the last 5 years, on and off, but I still don't feel competent. I have switched to MS Word, a couple of times, and switched back again, for reasons that barely make sense to me today. So, for me, at least, the right choice is far from obvious.

• I think this is a nice list though I would argue for (1) I wish the styles were really an alternative but they are broken (and confirmed by those MS experts on those forums) and (2) TeX is using an optimization algorithm for a very tough (NP-hard) problem and Word is not even trying (that's why TeX compiles, Word shows) and those tweaks you mention are typographical nudges when there is no good solution by the algorithm. However, it depends on the use case. If one is happy with the end product of Word I don't think anyone has the right to argue that TeX is better to a Word user. Dec 29, 2014 at 2:09
• But I can't agree with number 3. That's not correct in both typesetting part and also the font property part. The problem of math in Word is that it doesn't understand the context. It most of the time assumes and often does wrongly. Dec 29, 2014 at 2:12
• @percusse -- On the typesetting topic: a few details here: blogs.msdn.com/b/murrays/archive/2006/09/13/752206.aspx. I don't know what you mean by "doesn't understand the context". If the material is in a Math Zone, then it's math ?? Dec 29, 2014 at 2:29
• I'm semi-aware of those improvements but they don't show up in the end product at all or only optimized for simple cases. Even Leslie Lamport is working for MS so I'm not implying that they are not improving. Quite the other way around but there is no way that there is a competition yet. And for reemphasizing if you don't feel uncomfortable with the end product, nobody should force you out of MS Word. So it is a matter of satisfaction with the exception probably the math part. By context I mean the order and the grouping of symbols change behavior very much like ligatures in normal text. Dec 29, 2014 at 2:55
• regarding number 3, unless it has changed recently, the algorithm in word for formatting text is not based on the tex algorithm. with respect to math, it definitely is, and microsoft is responsible for designing and implementing the opentype math table, in which they even corrected an oversimplification that knuth made. so potentially, the math layout (with proper fonts) could be better with word. but it's not faster once you have mastered the names of the symbol commands. mouse clicks are demonstrably slower. Dec 29, 2014 at 14:15

I'm not sure today's zipped XML in Word is a step up - because only Word reads it reliably.

A colleague wrote her thesis in Word - now, given the colleague has no intention of paying for Word, Libre Office is a natural alternative. However the formatting is utterly and completely broken... Now you may blame this on Libre Office not reading the docx specification, however their point is that the specification exists mainly as an excuse as it is needlessly complex covering thousands of pages. Interestingly, saving said Word document from Word in odt, the open document format causes the same issues - breaking the entire formatting.... (Odt would however be read in flawlessly in Libre Office.)

Incidentally, formatting can also be problematic across much simpler documents once they contain a bit more than plain text... (I also cannot be bothered to pay for Word and use LaTeX exclusively, which means any forms in Word from my university end up being mangled on my laptop in Libre Office...)

LaTeX does not have such issues, except that some packages may be removed which may mean that every few years more "exotic" input may be unsupported. However, as others have pointed out correctly, the shelf life of LaTeX documents is very long as the basics have remained virtually unchanged for many years now. (While Word has gone through multiple iterations.)

• The XML word format is simply a blob in an XML wrapper, rather like saying that UUE files are plain text. Dec 25, 2014 at 11:09

There is one advantage left which has not been mentioned (or at least I have not read). As *tex files are just text files (which leads to a lot of advantages, see the other answers), you can use your favourite editor.

That may not be important for all users, but if I have to type a lot of text, I will love to that in Vim rather than in Word: (After "some" learning and configuration) this approach is a lot faster and way more convenient.

On the other hand, if you do not like Vim, you can use your favourite editor as well. That is the advantage here. We could even work on the same document while both using our favourite editor.

Edit: As mentioned in the comments, I did not really answer the question. This not not about "Word vs. LaTeX", but about the improvement of LaTeX since 2000. (After reading all these answers, I might have forgotten the actual question. Sorry for that.) However, my point is still valid: You get all the improvements of your favourite editor.

Vim 7.3 was released 2010; in 2000 only 5.7 was available. I do not even use Vim long enough to know a lot of improvements made in 14 years, but I guess I would miss many of them, if I had to switch to 5.7 right now. Of course not only Vim itself improved, there are also more and better addons/plugins available. For example: LaTeX-Suite (which I do not use, but that is not important here) exists since 2002.

• It's not about why TeX is better (though I agree). It's about what is extra for TeX recently. Dec 24, 2014 at 18:40
• @percusse: Oh, you are right of course, thanks. I made an edit to fix that problem.
– Keba
Dec 24, 2014 at 23:38

Support for PDF and press-ready printing, including colour models other than RGB, vector graphics beyond the simplistic (E)WMF --- these options have been available in (La)TeX almost since its inception.

• Whose inception? TeX goes back to 1977, LaTeX to 1983, while the first specification of .pdf is 1993. Dec 19, 2014 at 17:02
• i believe that earlier computer-to-plate software used postscript, which existed before pdf. before that, tex-to-typesetter, at astonishingly high quality, worked almost from day 1, with no pasteup; from there, a camera was needed, and real negatives. Dec 19, 2014 at 20:26

The question specifically asks for improvements made after 2000, but framing things that way makes it hard to point out the basic benefit of LaTex, one that hasn't changed since it was created nor is MS Word ever likely to 'catch up': LaTex is for writing code, Word is for writing (mostly) inactive text.

Having read so many competent and technical answers, this seems to be a completely obvious statement to the point of being trivial, but the differences are quite profound. One recent example to illustrate that basic point. Using \automultiplechoice package (AFAIK created after 2000) I can pass on the series of commands to the computer that will copy question sets from specified files in my computer, combine randomly chosen questions from each file in specific ratios, shuffle the questions as well as answer choices, then compile any number of individualized exams with active answer fields that can be easily processed with the help of a scanner. I can do this because using LaTex means I am writing active code.

Using MS word I may be able to write that code too, and then ... nothing. MS Word was never meant to process and compile code and that fact is not likely to change. There is no MS Word specific computer language that will enable that kind of functionality. I would have to type out 75 permutations of my 6 page exam manually, and score it manually as well.

• Well actually there is and many non-tech companies hire people for that. And that is, with the most disappointed voice, Visual Basic and macros. Not saying it is in par with LaTeX or anything but it is there. Mar 21, 2015 at 16:33
• @percuse Interesting. I am out of my depth here, as I don't use Microsoft products much, but I was under impression that Visual Basic was a separate piece of software to MS Word, and that it costs around \$500 for a full version. Is there any specific relationship between MS Word and Visual Basic? I understand there is a free version as well, perhaps in relation to MS Word that could be considered in the same vein as various packages that extend LaTex functionality? Mar 22, 2015 at 14:13
• Ah no. I think you are confusing VB with VBA because of my mistake. I meant VBA (Visual Basic for Applications). Many firms who do extremely repetitive tasks (forms, spreadsheets etc.) make document templates using an API within MS Office to VB and that framework is called VBA. For example imagine you have a document that you would like to export automatically to some other format then you can program a macro to fetch and arrange certain parts of the document etc. I'm quite happy to realize that I don't remember anything about it:) microsoft.com/en-us/download/details.aspx?id=40326 Mar 22, 2015 at 18:27
• > I don't use Microsoft products much. Then it's odd that you make comments about their deficiencies. Your ideas about Word APIs, available languages, and costs are all incorrect. Mar 28, 2015 at 6:14
• Okay then, perhaps you'd care to educate? As for the extent of my experience, I have used Powerpoint presentations with Macromedia content and some macro-enabled Excel spreadsheets. It strikes me that it's only used for routine, repetitive tasks. If there is a way to use MS Word to compile more complex code, perhaps it would be useful to provide that information here, I am genuinely interested. By the way, I have not mentioned any perceived 'deficiencies' in my post, my point is that MS Word and LaTex have a different scope, hence the use of quotes around 'catch up'. Mar 28, 2015 at 11:52

The improvements to LaTeX since the year 2000 for me is the LaTeX editors are much more sophisticated today and I do believe that we have a wider choice of editors.

Also, thanks to the generous people who have created instructional videos and shared them on YouTube, and to websites like this one, the learning curve has diminished.

I agree with the post just above this one, yes it is the speed of the computers that has helped (probably that is what drove the improvement of the editors)

Speed! Quite simply, (La)TeX has become much more useable with faster computers. From the 80286 I used in 1986 to my present Macbook Pro, this is likely where I have seen the most progress. I Imagine Word benefited as well, but for TeXing the improvement has been huge.

The next big jump I am waiting for is a rendering engine that will show changes in real time even for long and complex documents.

• I doubt you'll see real time updates without a fundamental architectural change in latex unfortunately. Dec 30, 2014 at 1:09

The major improvements in LaTeX is that people have been writing plugins to do useful things. TeX and LaTeX are still viable for this reason, and will be for many years to come.

Word has no open API to allow developers to add functionality, and you can hit an insurmountable brick wall at some point or another.

On the other hand, I opened up LaTeX documents created under EMTEX for OS/2 + Semware for DOS + rexx scripts, and a series of DOS editors, into Windows TeXLive + WinEdt, and although the DVIs did not work after 15 years of technological improvements, the new ones compiled flawlessly.

Documents I created under MultiMate, or AmiPro, I had problems importing into Word, because Word lost all of the hierarchy, and MM was not a big player, it is hard to get stuff to run on the dusty deck.

So in effect, LaTeX got it right.

Here is somewhere it could be improved:

When you set up a distribution, you should be able to store the improvements in a single directory, and simply reload these. For example, my main box was overheating, and I started using the ThinkPad. You can easily set the distribution up, and because I make a special effort to keep the additional upgrade/extras in a directory, it was fairly easy to update it to full working order without having to grope and search for stuff.

• > Word has no open API -- Yes it does. And there are many add-on packages created by third parties. Not as many as there are LaTeX packages, though. Dec 28, 2014 at 7:29
• I don't understand the last paragraph. Why was special effort required? System-wide non-distribution stuff: TEXMFLOCAL. Personal non-distribution stuff: TEXMFHOME. That's part of the standard TDS hierarchy (at least as implemented in TeX Live). Maybe special effort is needed in MiKTeX?
– cfr
Dec 29, 2014 at 2:01
• If you try to set something up on a new laptop, you can rebuid from a base install, additional and updated stuff. I had to set up a working distribution from TeXLive 2003 + what i thought were updates. It missed some things, such as the updated english file for babel. I got most of it right, though. Dec 29, 2014 at 4:29