
Do we need a WYSIWYG editor for Tex, LaTex, and AmsTex?
Tags: Latex, Lyx, Scientific Word, TexPosted in Technical (ms word, tex), Tips, useful software
I still remember in the 1980’s how impressed we physicists were when we discovered Tex. The program was written by Donald Knuth. The macro package Tex is so good and complete that all new developments are mere front ends and user interfaces to Tex, of which Latex and AmsTex packages are the most popular. Newer distributions deal with newer hardware, new fonts and better font management, and pdf creation, but the fundament is still Tex.
Those scientists, like chemists and biologists, that use an occasional mathematical formula can do without Tex. All kinds of handy add-ins allow incorporating math formula’s in standard office documents. However, if your paper has many math formula’s the Tex-way is the only solution. In the rest I will limit myself to LaTex.
Compiling
A typical Latex cycle is a source code ascii file (extension usually .tex) that is compiled by a “latex” program into a dvi (device independent file) that subsequently can be viewed or printed. The learning curve for LaTex is quite long. Opponents of the Tex-approach always complain about the lack of a WYSIWYG (What You See Is What You Get) editor. They are used to MS Word, or alike, with a powerful Graphical User Interface (GUI). I am not going to start a discussion here about whether or not WYSIWYG and GUI’s is the best way. My opinion is that I agree with the Unix world that a graphical user interface and a WYSIWYG approach is in general inefficient for experienced users.
Math from a scratch?
A typical LaTex source file looks horrible for inexperienced users. To give an example:
\[ \int_0^\infty \sum_{l=0}^\infty\frac{ A_l ({\bf x})}{2 \pi}\]
will generate the formula:
Some of my students really develop math using the LaTex. I find this very difficult. What I often do is: use a pen write math, correct it, and correct it, and correct it, and then put it in LaTex.
If we would have a good WYSIWYG editor we would be able to develop math immediately from scratch into a usable tex file.
Requirements for a WYSIWYG LaTex editor
- The GUI interface should be user friendly, and compiling and printing should be transparent to the user. The help system should focus on how to use the editor, not on explaining LaTex, because enough good documentation exist for that.
- An acceptable WYSIWYG editor would have to be backward and forward compatible: that is to say it should be able to import LaTex files from any old or new Latex version.
- It should be able to export clean LaTex files, that is without relying on macros not being part of the standard LaTex distribution.
- If not open source, the licensing should be reasonable
Requirements 2 and 3 will allow authors to switch between any editor they like. Some co-authors might want to use the WYSIWG interface, others might want to use of the raw ascii interface.
The only two WYSIWYG editors for LaTeX I know of are the open source Lyx and the commercial Windows program Scientific Word sold by software company MacKichan.
Lyx
Lyx is open source with a Unix taste, and as it is free users should hold back with their complaints. Well, Lyx is horrible. At least on Windows. In my case the install procedure hung time after time on missing LaTex packages. The documentation is awful, unclear, scattered, inconsistent, ugly .. The printing of a document is terrible. Apparently it helps when you install Cygwin, a Unix-environment for Windows. So Lyx violates my requirement 1. It also violates requirement 2, as I tested it with at least ten bona-fide LaTex files, that were accepted by scientific journals in the past. In all cases Lyx could not handle therm and told me there were fatal errors in them. The Lyx people advise to write papers specially for the Lyx system and indeed Lyx stores the Latex information in a non-Latex file. Horrible.
Scientific Word
Scientific Word is much better than Lyx. Requirement 1 is fulfilled and requirement 2 is also fulfilled. Requirement 3 is only partly fulfilled. Rather than removing all SciWord stuff when cleanly exporting, it comments their own directives out in the source file. In addition it rearranges the LaTex original file. Moreover users that import a file, do not change it, and export it again will discover that it does not compile anymore because the proprietary file tcilatex.tex is needed. This is the wrong way. The SciWord developers should have developed a standard Latex package, perhaps call tcilatex and made it part of any Latex distribution.
I have uses SciWord a lot, but I am about to abandon it because of its licensing conditions. It is way too expensive, $525 for academic use and $180 for students. Happily one of my affiliations has a site-wide license. However, the licensing scheme is cumbersome. It is connected to one computer and it is per year. I use it on four computers. So every three months one is expiring, without issuing a warning. It always happens to me in a weekend or on a conference and then I am out of working program.
My Solution
My solution is the fountain pen again. I write my math. After it is done I use the WinEdt ascii editor with the MixTex distribution as backend. It is fast and robust.
9 Mar 2010 23:47, cpbotha
(I’m typing this comment for the third time now… *sigh*)
Many people don’t know this, but Google Docs has a built-in equation editor that natively uses LaTeX! You can build up your equation with the WYSIWYG functionality and it updates the corresponding LaTeX in real-time, or you can enter LaTeX directly, and it updates the typeset math in real-time. See this screenshot: http://cpbotha.net/wp-content/uploads/2009/09/google_docs_ss1.png
Personally I type math equations directly in LaTeX, but the Google Docs equation editor is very useful if you quickly need an equation typeset, or you want to teach someone else how to use LaTeX math. One could also use it in addition to a normal text editor, purely for doing the math bits in LaTeX.
10 Mar 2010 11:07, Jacopo Bertolotti
I’m aware that LaTeX is strongly unfriendly with newbies and I would never recommend it to anyone who has just to type a commercial letter or anything like that. Yet, once you overcame the steep learning curve, everything just flow smoothly and the degree of control you have on the final result is infinitely higher than any MS Word-like program will ever allow.
As an example my wife recently typed her Master thesis on international adoptions (zero equations and just one graph) in OpenOffice Write. Once she finished I took it and retyped everything in LaTeX giving it a decent (and consistent from beginning to end) format, a decent bibliography and so on. The final result was just so much better than anything you can make with Word-like programs.
That said I totally agree that a decent WYSIWYG editor would probably help making TeX a standard outside the Math-Physics community.
I’m not aware of any good editor that satisfy all three criteria, but if the main requirement is a help for typing equations then I can suggest the use of http://www.codecogs.com/components/equationeditor/equationeditor.php?lang=en-en . It requires no installation at all (you access it through your browser) and allows you to see in real time the appearance of the equation you just typed. As a (big) bonus you can export the equation as a jpg/png/gif/svg/pdf image and use it in your slides for a presentation (I do it systematically since I discovered it).
11 Mar 2010 13:30, Ad Lagendijk
Charl,
thank for the comment. I hope your typing in your comment was not related to problems with this site. We have installed recently a number of filters as we are flooded with comments by spambots.
–
There are quite a number of on-line real-time LaTex “translators”. Google Docs is one f them, and Jacopo is pointing out another one in his comment. They are indeed very good for didactic reasons, and they can be used for generating bitmaps in presentations. For the latter I use TexPoint.
11 Mar 2010 14:03, Jacopo Bertolotti
TexPoint is nice but it is non-free (ok, 30 dollars are not a huge amount of money) and, sadly, don’t work on Linux.
12 Mar 2010 22:05, Jan
I agree that the Windows installation can be tricky, it worked much better on the Mac or Linux. But one tip: try using the complete package of the alternate installer: ftp://ftp.lyx.org/pub/lyx/bin/1.6.5/LyX-165-4-22-AltInstaller-Complete.exe
It installs the whole environment and all the bindings. Always worked well for me.
–
But reading your comments again I realize that my requirements are somewhat different. I rarely use LyX to import a LaTeX source. After all, that is already done and dusted. Instead I use LyX as a much more comfortable alternative for writing a paper from scratch. Depending on the set of colleagues, I might have to use Word or LaTeX. If we use LaTeX, I tend to use LyX with the Revtex4 template it ships with and it works perfectly fine, including most importantly the final export to Latex.
–
But as somebody already commented: Google Docs is looking like a nice alternative for collaborative paper writing since it supports native Latex equations. One important tipp for that: when you export a Google doc to a text file, it actually exports the equations in plain latex! Nice!
–
Thanks for your interesting blog!
15 Mar 2010 9:59, suzan
I don’t agree with this:
Those scientists, like chemists and biologists, that use an occasional mathematical formula can do without Tex.
For formatting papers according to the stylesheet provided by journals and conferences, LaTeX is indispensable.
The MS Word stylesheets provided by conferences are often very unprofessional and lead to a lot of struggling.
With LaTeX, you can simply attach any stylesheet to the same paper to make it suitable for a specific journal or conference.
Let alone the bibliography! How would you manage your references if you didn’t have bibtex?
15 Mar 2010 10:35, Ad Lagendijk
@Suzan,
thank you for your reaction. I am a chemist by training myself. There was nothing derogatory meant in my post. I agree that Latex produces superior output. And for physicists and mathematicians it is a must. But for some reason many scientists insist on using Ms Word in combination with EndNote for instance. I wish every scientist would use LaTex. There is even quite a threat for the LaTex world: MathMl. Rather than using LaTex as a base the wheel is reinvented for showing math formula’s on web pages.
15 Mar 2010 10:51, Jacopo Bertolotti
@ suzan: Sadly a significant fraction of journals that focus on chemistry and material science (e.g. Advanced Materials) accept manuscripts only as MS Word files.
15 Mar 2010 11:21, suzan
@Ad I was not offended by your point; I just didn’t agree 🙂
@Jacopo: That is really sad! In my field (computational linguistics), there are some conferences/journals that only accept LaTeX! Most of them provide sylesheets for both word and LaTeX.
15 Mar 2010 20:41, Klaas Wynne
Sorry but I am a physicist (although a chemist by training with the same PhD advisor as Ad) and I would rather have my nuts pulled off with a blunt instrument than using La-bloody-Tex. However, if some people prefer to use 1970’s software…good luck to them. To me, LaTex prepared documents look positively rubbish with (typically) poor spelling and grammar because of the lack of easy to use spelling and grammar checkers. I would urge any scientists (physicist or otherwise) whose first language is not English (say, Dutch, to name a random example), to use MS Word and to pay careful attention to the recommendations made by the grammar/spelling checker. In fact, that advise would extend to British and American writers as well, as their English tends to be quite poor as well. I remember my dad having Donald Knuth’s book on his book shelf. Knuth designed Tex because in the 1960’s there were no other options. It is now 2010. People do no use punch cards anymore. Sane people also do not use Tex.
15 Mar 2010 23:43, Paolo Scalia
I agree that Latex is not user friendly and even discouraging at times. I have to say though that not having to worry about numbering my equations, sections, subsections, positioning my figures and updating the table of contents is a big plus. If in combination with that a solid way to handle references (like jabref) and a nice stylefile are used, then I would prefer Latex over MS Word. Furthermore WinEdt has a decent spellchecker.
In my opinion Latex’s true drawback manifest itself when one needs to write many long mathematical expressions. In all honesty, having to write involved and long mathematical expressions in Latex directly is a pain and drives me berserk.
I have had a rather positive experience with Lyx. At the moment I use it for the most mathematically dense parts of my thesis. Once the Lyx file is complete and I am happy with it I ask Lyx to generate the correspondent Latex file. I then paste this Latex file into my thesis Latex file, right where I need it. It works like a charm. The only annoyance I have found is that the formulas in the Lyx generated Latex file are a bit cluttered. Of course you don’t see that when you compile the Latex file.
The other way around (from Latex to Lyx) also works but not as smoothly, I found. I believe this is mostly due to me not having found the right settings or downloaded all of the needed
packages.
16 Mar 2010 8:26, Klaas Wynne
@Paolo: My point is that with Latex and all its associated tools you have to go through multiple steps and procedures to get a result. It sounds like you have to copy a “formula” from your Latex source file to Lyx, edit it there, save to another file, open that, copy back into your Latex source file, compile, check that the result is OK. The procedure in Word+Endnote+MathType (WEM) is: double-click. This means that you have time to check your work and to think about what to write. WEM does TOCs, equation numbering, section numbering, styles, and everything that has been mentioned before. If the Latex people keep maintaining that Latex produces superior output then I think it is time for a proper contest. Anybody up for that? What if we reproduce a paragraph with lots of maths from a recent PRL? For proper comparison, we should also keep track of the amount of time required to reproduce that paragraph. Ad could be the referee.
16 Mar 2010 10:15, Jacopo Bertolotti
@ Klaas: I think that the difference in perspective between LaTeX lovers (who usually are Word hater) and Word lover (who usually are LaTeX hater) is a bit too deep and personal to be the subject of a fair contest.
Word lover usually rate very highly anything that can be done with just a few clicks of the mouse while LaTeX lover don’t care so much and are more attracted to anything that can be finely controlled with a couple of code lines. This is something that pertain to personal taste more than to some objective and measurable parameter.
Personally I hate any equation editor that requires me to click around like crazy. I find much faster and comfortable to just type without ever leaving my keyboard but I know that different people might have different feeling about that.
If you prefer WYSIWYG over WYSIWYM editors maybe LaTeX is not the best choice for you. Maybe (and here I’m talking to Ad and Paolo) you would feel better with a decent equation editor for Word (or OpenOffice or your text editor of choice).
To conclude just a few remark on LaTeX: any decent editor for TeX/LaTeX have a built in spell checker on the same level on what you find in Word. At the same time the “multiple steps” you need to obtain a result in LaTeX amount to click once (or use the keyboard shortcut to be faster) on the “compile” icon in your editor and once on the “view dvi/ps/pdf” icon to actually see the result. Notice that the latter step can be done just once and for all since most editors are smart enough to update what you are visualizing each time you compile your LaTeX source. The times when you had to go to the shell and type some esoteric commands to get anything are long passed.
29 Mar 2010 9:23, Maarten
About LyX:
I have been using it now on Windows and Linux for more than 2 years and for me it works perfect, and I am afraid that Ad has not given it a real chance. I’ve never learned to write complete documents with a LaTeX editor, but I am familiar with the capabilities of TeXnicCenter, just to but things in perspective. I found LyX to have learning curve that is not too steep and much of the questions are you’d possibly have are answered in the help files that come with it or in the LyX wiki. I found the developer community to be fairly busy and helpful, and that much has been improved over the years, with every new version.
–
I’d like to elaborate on the 4 points that were raised in reverse order:
–
Point 4: LyX is open source, free and platform independent; although my experience is that a MikTex distribution works best. For Windows, the AltInstaller is clearly the best option, and if MikTex is allowed to update packages on the fly, then there should not be any problem with using it.
–
Point 3: LyX can export to plain LaTeX and I’d recommend using LyX for this purpose, as also explained by Jan. I’d like to add that simple tables can just be copied from Excel into LyX, all possible tweaks like figure positions, caption mark up etc. can easily be tried out before using it in the ‘plain’ editor. REALISE: this can be done without all the open and close brackets and code lines. Besides: the user can choose how to write the mark up: mouse clicking, hot-keys or writing TeX code. Also it is easy to change the document structure on the fly (changing section order or level). If one copy and paste’s an equation from LyX to some textfile, the plain LaTeX is copied, such that is very useful as an equation editor for TeX, instead of using mathtype, or Google Docs.
–
Point 2: I have found LyX supports also olderLaTeX packages, but also that some journals use obsolete classes, or have tweaked classes that they should not. This has been the case for me with the copernicus class of the European Geophysical Union (so indeed, I am not a physicist). If LyX is used as an extra editor for equations and other heavy ‘coding,’ one will not have any problems with this. The main advantage of LaTeX over Word lies here as well: one can always see and edit the ASCII file and there are no problems with backward compatibility, differences in system language or dependency on other third party software like Endnote or Mathtype.
–
Point 1: I agree that the help system is not always too clear, or that an extra internet search into the wiki or mailing list may be needed for fancy stuff. If one has done the 1 hour tutorial, one can at least use LyX for creating nice LaTeX documents and his own templates for making the PDF.
–
To conclude, some remarks on other things raised in the discussion put in the LyX perspective: LyX is good for collaboration using a track changes system, and different methods for making comments. Also the spell checker is good, although not yet on the fly like in word, and grammar is not properly evaluated. The upcoming LyX 2.0 will be better in this respect. Last but not least, I really like the easy way one can make his own template for writing a report for example. But remember, just see it as an extra editor that is just much easier to use than a plain text editor!
15 Apr 2010 13:08, PM
I do not agree with you about LyX. Even I hated it initially, since I liked typing the LaTeX code. But now, I have switched over to LyX completely (both in Win and Linux), because of its easy to use crossreference, citation and table properties. No doubt it is not the greatest, but its still free and very functional. And if one is familiar with typing LaTeX equations, its equation editor is really amazing with the complicated stuff,as one can type code for the simple parts and click on buttons for the horrid stuff like fractions and bracket nesting that swell up latex code.
Also, use zotero, which unlike endnote is both free and exports to bibtex.
P.S. I am from a business school, where LaTeX is hardly popular 🙂
2 May 2010 15:49, Vasilij M.
Just for information …
You can also see such WYSIWYG LaTeX editors as
BaKoMa TeX Word
Scientific Author
MicroIMP (probably stalled)
22 Oct 2010 17:43, Darren Ardville
The learning curve of TeX is simply not worth it. I entered graduate school in mathematics in 1988 at Rutgers University in Piscataway, New Jersey. Luckily, not a single one of my graduate math courses required any typeset material. Thankfully, the professors all had the sense to focus on the math – the pure content – and not to force students to waste years of their lives chasing down obscure commands in TeX. Even my math Master’s Thesis is entirely hand printed.
I first encountered Design Science’s Mathtype in 1997 and have never looked back. All the papers and documents that I prepared for all the math courses that I taught at community colleges (Mercer County Community College, Burlington County College) and four-year colleges (Monmouth U, Rutgers), both fulltime and parttime, I did with MS Word & Mathtype.
I typeset my entire 275-page doctoral dissertation for Rutgers (PhD 2000) in MS Word & Mathtype.
I have published 5 papers in peer-reviewed math journals – 7, if you consider arxiv.org “peer-reviewed”. And I am continuing to do research – all thanks to MS Word & Mathtype.
I have no idea how mathematicians get ANYTHING done in TeX.
It is has been 20 years – roughly since 1990 – since I was made aware of TeX and its numerous variants.
I know that MS Word & Mathtype HAD a translation option which allowed me, with ONE click of the mouse, convert my entire document at once into TeX. I did that on about 2 or 3 occasions for the benefit of some mathematicians and journals. But, I had no way of viewing the horrendous mess that was output.
For some reason, I cannot find that option any more. I seem to be forced to convert only 1 Mathtype object at a time into TeX, so at present, it is not worth my time doing the conversion, with thousands of Mathtype objects in a typical 20-page document.
I need a lot of help on my math research. Unfortunately, places such as mathoverflow.net make it impossible to post a question, as they require the stupidity and insanity of posting in TeX.
There is no logical reason I could not upload a PDF of my MS Word document to them.
So, yes – it has been 20 years since I’ve known about and off-and-on tried TeX, and in that 20-year learning curve, it is not yet possible to produce even a single line of useful meaningful mathematics.
15 Dec 2011 6:52, Alice the Intruder
Thank you for this post! You reminded me of yet one more reason why I LOVE Mathtype by Design Science and like Microsoft Word 2000. And that reason is that I can actually DO MATH with MS Word and Mathtype. MS Word 2000 certainly has its enraging faults. Nevertheless, combined with Mathtype 6.2, I have published 5 peer-reviewed math papers, 7 if you include arxiv.org as “peer-reviewed”, a 275 doctoral dissertation, and countless papers for the math courses I have taught.
There was a reason WYSIWYG was invented: so users could USE the computer to put out documents on other subjects, rather than spend decades researching the tool itself.
By contrast, since I first heard of TeX and its numerous hideous mutant variations in 1988, I have attempted to use it on around 5 different occasions. I never got past downloading the free source code. The utter worthless impracticality of typing code from essentially the 0s and 1s of machine language proves that TeX users don’t actually want to do math, but prefer to waste eons searching the ends of the galaxy for the exact command to make a subscript.
Many times, when I cannot get all my math ideas organized on paper, and when formulae become too cumbersome to do by hand, and I need to copy and paste formulae into other formulae flawlessly, I do these procedures in MS Word and Mathtype. So, the Mathtype and MS Word do not ONLY allow me to TYPESET my document. They actually allow me to DO THE MATH ITSELF!
30 Jun 2012 17:06, Lee A. Young
I am surprised no one mentions BaKoMa, a very good LaTeX editor and creator offering full WYSIWYG display. Visit
http://www.bakoma-tex.com.
Lee A. Young
21 Sep 2012 4:59, G.A Betancourt
i’m an engineer, and like the simplicity of latex, once you have learned how to do properly, typesetting a document is not difficult.. i find lyx obscure and difficult to use, mainly because you cannot see the underlaying code. i seen people of my university sitting hours trying to catch lyx document errors, same errors in latex are easy to find.
13 Mar 2013 23:36, MJ
I think latex is great in general but it can be cumbersome sometimes. For example, I wrote my thesis using latex. When I was proof reading it and found some small mistakes or typos, I had to go back and forth in the long source document just to find out the locations of the typos. It would be nice if I could just make changes in the rendered version and the editor could change it in the tex source for me, like BaKoMa can do.
7 May 2013 15:09, Daniel
It’s true that WYSIWYG are definitively annoying for experienced users. On the other hand, when I was a second year undergrad most teachers started requiring that we typed our homework (using LaTeX). Although some professors were against WYSIWYG software (some even called it “butterfly LaTeX) it helped me learn LaTex quickly than it would have been if I had sat down with a manual trying to type for scratch. At one point I realised I wasn’t using any of Lyx’s add-ons to insert equations and I wanted more control over the way my documents were presented, so I took the leap and move on to LaTex. I still use LyX when I want to type a quick documents (in particular, quizzes or short tests for students).
5 Jul 2013 7:28, Daniel KIan Mc Kiernan
My formulæ are often complex and may need multiple revisions; and I work far more comfortably with visual tidiness. A WYSIWYM or WYSIWYW or WYSIWYG editor therefore accelerates my progress.
I use LyX (on Linux computers) as I am developing my work, and then cannibalize exported LaTeX files to produce something ready to submit to a journal.
I tried Scientific Word (as part of Scientific WorkPlace) back in the mid-’90s, and my experience was ghastly. Short of being compensated (with accumulated interest) for the many hours of lost time, I wouldn’t consider giving it another chance.
The formula editor bundled with MS Word simply could not handle the nesting in some of the expressions that I wanted to use. Expressionist did better, but I still bumped-up against a ceiling. I did not have this problem with the formula editor of OpenOffice, but a problem of font-fallback proved catastrophic. I would note, though, that Writer2LaTeX did a fine job of converting OpenOffice formulæ.