Article suggests researchers should be prevented from producing documents in LaTeX

[tsd]

"Given these numbers it remains an open question to determine the amount of taxpayer money that is spent worldwide for researchers to use LaTeX over a more efficient document preparation system, which would free up their time to advance their respective field. Some publishers may save a significant amount of money by requesting or allowing LaTeX submissions because a well-formed LaTeX document complying with a well-designed class file (template) is much easier to bring into their publication workflow. However, this is at the expense of the researchers' labor time and effort. We therefore suggest that leading scientific journals should consider accepting submissions in LaTeX only if this is justified by the level of mathematics presented in the paper. In all other cases, we think that scholarly journals should request authors to submit their documents in Word or PDF format. We believe that this would be a good policy for two reasons. First, we think that the appearance of the text is secondary to the scientific merit of an article and its impact to the field. And, second, preventing researchers from producing documents in LaTeX would save time and money to maximize the benefit of research and development for both the research team and the public."

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0115069

[dnj...@gmail.com]

1. In my world (engineering) is the writing time of the final article probably less than 1% of the total time spend on the research, data acquisition, data analysis, simulations, etc.

2. The power of LaTeX is with books/theses with hundreds of references, publication quality graphics and tables. The ease of changing and moving content around is still a LaTeX stong point

3. Integration with other tools such as Matlab/Python programs or CAD/CFD/FEM that generates graphics, tables and text, including the program listings. The full document can be updated with a LaTeX run or two. Not remotely possible with Word!

[Jack Ryan]

> "Given these numbers it remains an open question to determine the amount of=
> taxpayer money that is spent worldwide for researchers to use LaTeX over a=
> more efficient document preparation system, which would free up their time=

> to advance their respective field.
>

> http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0115069

The most obvious oversight is that the article neglects the value of
information acquired as a user advances.

The struggles a LaTeX user faces are primarily the result of SUEs
(stupid user errors), in which case they learn something of value that
pays off in the long run.

The struggles a Word user faces is the tool that tries to guess what
the user wants, not doing what the user expects, ultimately due to a
bug in the tool (or format). The knowledge they acquire in this case
is bug circumvention, which is a waste in the short-term with no
long-term payoff (iow, the bug will eventually be fixed and replaced
with different defects). It's not a buiding-block of knowledge that
can be used to advance other situations.

The article crudely compares the raw amount of "waste", failing to
acknowledge that the *overhead* in composing Word documents has a
higher waste/value ratio. Nevermind the societal detriment of feeding
a commercially interesting proprietary lock-in, directly opposing
academic insight and hindering progress.

[Phillip Helbig (undress to reply)]

In article <aaec7d46-c1dc-4948...@googlegroups.com>,

dnj...@gmail.com writes:

> 1. In my world (engineering) is the writing time of the final article
> probably less than 1% of the total time spend on the research, data
> acquisition, data analysis, simulations, etc.

Right.

> 2. The power of LaTeX is with books/theses with hundreds of
> references, publication quality graphics and tables. The ease of
> changing and moving content around is still a LaTeX stong point

Right. But it is worth it for articles as well. It's not just LaTeX
itself, but also BibTeX.

> 3. Integration with other tools such as Matlab/Python programs or
> CAD/CFD/FEM that generates graphics, tables and text, including the
> program listings. The full document can be updated with a LaTeX run or
> two. Not remotely possible with Word!

Right. I use PostScript figures generated by Fortran!

[Ulrich D i e z]

tsd quoted:

> In all other cases, we think that scholarly journals should request authors to submit their documents in Word or PDF format.

So let's submit documents in PDF format, produced via pdf(La)TeX...

Why focussing on "Word or PDF format "instead of focussing on
something based on SGML resp. XML -- e.g., DocBook or DITA?

( Afaik there exist tools both for producing LaTeX code and for
producing Word documents from DocBook... )

Ulrich

[Robert Heller]

And I wonder how much M$ *paid* the researchers... I note that the article
makes no mention of OpenOffice / LibrOfice, only *specifically* MS-Word.

>
>

--
Robert Heller -- 978-544-6933
Deepwoods Software -- Custom Software Services
http://www.deepsoft.com/ -- Linux Administration Services
hel...@deepsoft.com -- Webhosting Services

[Robert Heller]

An additional observation about the PLOS ONE article: it has some of the same
'flavor' as the anti-passenger train crowd (mostly coming from the
Oil/Auto/Airline industry supporters/lobbyists). The *numbers* suggest that
air travel is faster, inter-city buses are more fuel efficient, and
automobiles provide greater freedom (when and where to travel). Despite these
*numbers*, people keep riding trains (and Amtrak' ridership has increased
*every* year since its creation). The anti-passenger train crowd cannot
explain the (growing) popularity of passenger trains. Like LaTeX, users of
passenger trains like their chosen mode of travel, despite the *numerical*
reasons that they shouldn't (eg higher cost, lack of flexibility, longer
travel times, on-time performance problems, etc.). This makes me wonder about
the objectivity of the authors. They may *claim* that they are not in the pay
of Microsoft and have no bias, but I still wonder. Especially given that
paragraph about seeking to change people's 'beliefs'. That is a little
disturbing (can you say 'Thought Police'?). Also I find some of their
measurements a little questionable. There seems to be a quantity vs quality
measurement. I wonder if the MS-Word using researchers are just more wordy
(more 'words', not necessarily more actual useful content) than the LaTeX
using researchers. Eg. computer science and mathematics are more formula and
equation heavy and social sciences are more textually based. Computer
scientists and mathematicians are more likely to use LaTeX and social
scientists are more likely to use MS-Word. I wonder if the article's authors
thought to make some sort of adjustment for these biases.

The article suggests that LaTeX has a steeper learning curve and imposes
stricter formatting 'rules' (my guess the 'formatting errors' are due to
encountering LaTeX (and TeX) 'errors' -- stuff like mis-balanced braces,
misspelled command names, etc.), but even so, people seem to 'enjoy' using
LaTeX more than using MS-Word.

The article also seems to completely leave out the fact that MS-Word only runs
under MS-Windows, and that both MS-Word and MS-Windows are *costly* commercial
software.

[alan]

On Saturday, December 27, 2014 9:12:20 AM UTC-5, Robert Heller wrote:
> At Sat, 27 Dec 2014 07:12:22 -0600 Robert Heller <hel...@deepsoft.com> wrote:
>
> Also I find some of their
> measurements a little questionable. There seems to be a quantity vs quality
> measurement. I wonder if the MS-Word using researchers are just more wordy
> (more 'words', not necessarily more actual useful content) than the LaTeX
> using researchers. Eg. computer science and mathematics are more formula and
> equation heavy and social sciences are more textually based. Computer
> scientists and mathematicians are more likely to use LaTeX and social
> scientists are more likely to use MS-Word. I wonder if the article's authors
> thought to make some sort of adjustment for these biases.

And the result of this is that the MSWord users were just better typists? I'm assuming that's what you mean, because in the experiment everyone typed the same three pages, one of which included lots of math.

> The article also seems to completely leave out the fact that MS-Word only runs under MS-Windows,

This is simply not true: Word runs on the Mac OS too and has done for years. It doesn't run under Linux though although the Web version (http://www.dedoimedo.com/computers/office-online-linux.html) which is free, in fact, does.

Alan

[Alain Ketterlin]

tsd <t...@tsdye.com> writes:

> "Given these numbers it remains an open question to determine the
> amount of taxpayer money that is spent worldwide for researchers to
> use LaTeX over a more efficient document preparation system, which
> would free up their time to advance their respective field.

[...]
> [P]reventing researchers from producing documents in LaTeX would

> save time and money to maximize the benefit of research and
> development for both the research team and the public."
>
> http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0115069

This paper is beyond stupidity, so much that it is difficult to argue
against. If the goal is to save taxpayer money, I suggest preventing
people from asking meaningless questions (writing vs formatting),
developing bogus methodology (averaging novice and expert performance,
huge standard deviations on a population of 10), producing insignificant
results (a vague notion of "error", including orthographic errors for
vocabulary most participants were unfamiliar with), and drawing
excessive conclusions ("preventing researchers from [using] LaTeX" --
seriously?) to give the impression of substance.

And with a broken link (in both the online and PDF versions) in the
8-entry bibliography.

My guess is that it is a hoax.

-- Alain.

[GL]

I agree with M Ketterlin : this is a fucking hoax ! ;-)

I'm not sure as far as LaTeX is concerned, but TeX remains since
1982 (more than 30 years) the leading typesetting system for science,
and will for the next decades for sure.

[Torsten Bronger]

Hallöchen!

GL writes:

> [...]

>
> I'm not sure as far as LaTeX is concerned, but TeX remains since
> 1982 (more than 30 years) the leading typesetting system for
> science, and will for the next decades for sure.

Is there any solid evidence for this? I would be particular
interested in statistics about how many journals accept which input
formats.

Tschö,
Torsten.

--
Torsten Bronger Jabber ID: torsten...@jabber.rwth-aachen.de
or http://bronger-jmp.appspot.com

[Axel Berger]

Jack Ryan wrote on Sat, 14-12-27 06:04:

>The article crudely compares the raw amount of "waste",

And not even that. As I've often said before, I could easily live with
the quality of Office programs' output but the hassle of using them and
their horrible ergonomics drives me up the wall. It's the ease and the
useability that hooks me to LaTeX (and my favourite editor).

[Robert Heller]

At Sat, 27 Dec 2014 07:54:50 -0800 (PST) alan <alan...@gmail.com> wrote:

>
> On Saturday, December 27, 2014 9:12:20 AM UTC-5, Robert Heller wrote:
> > At Sat, 27 Dec 2014 07:12:22 -0600 Robert Heller <hel...@deepsoft.com> wrote:
> >
> > Also I find some of their
> > measurements a little questionable. There seems to be a quantity vs quality
> > measurement. I wonder if the MS-Word using researchers are just more wordy
> > (more 'words', not necessarily more actual useful content) than the LaTeX
> > using researchers. Eg. computer science and mathematics are more formula and
> > equation heavy and social sciences are more textually based. Computer
> > scientists and mathematicians are more likely to use LaTeX and social
> > scientists are more likely to use MS-Word. I wonder if the article's authors
> > thought to make some sort of adjustment for these biases.
>
> And the result of this is that the MSWord users were just better typists?
> I'm assuming that's what you mean, because in the experiment everyone typed
> the same three pages, one of which included lots of math.

Well, maybe people in the social sciences write more 'wordy' papers vs.
computer scientists and mathematicians write papers that are more heavily
leaning to math and equations. So, the social scientists might be better at
typing long streetches of just plain text.

>
>
> > The article also seems to completely leave out the fact that MS-Word only runs under MS-Windows,
>
> This is simply not true: Word runs on the Mac OS too and has done for years.
> It doesn't run under Linux though although the Web version
> (http://www.dedoimedo.com/computers/office-online-linux.html) which is free,
> in fact, does.

There are some interoperatibilty issues between MacOS versions of MS-Word and
the MS-Windows versions -- I worked for a trio of people heading the research
group I worked for. Two used MacOS and one use MS-Windows. When they
worked on a joint paper and traded the MS-Word .doc file things would often
get messed up. They eventually gave up and had the secretary do *all* of the
actual typing. She would then *print* three copies of the draft and the two
profs and the senior researcher would mark up chages with their red pencil and
the secretary would edit the file. All this 'high tech' software and things
were functioning on a fairly old-fashioned level. (Almost all of the grad
students were using LaTeX on Linux workstations.)

Also, I don't believe it is normally possible to get MS-Word pre-installed on
Macs, where most PC OEMs offer to pre-install MS-Office as part of the base
install. For *many* people with Macs, MS-Office is more often not available
than for people with MS-Windows PCs. To a certain extent, MS-Office implies
MS-Windows. And in the case of this article, only machines with MS-Windows
were used (with MS-Office/MS-Word). There was no mention of MacOS. (There
was mention of Linux, but I'm assuming that the people with Linux machines
were using LaTeX, since no mention was made of OpenOffice or LibreOffice.)

>
> Alan

[alan]

On Saturday, December 27, 2014 3:47:48 PM UTC-5, Robert Heller wrote:

> At Sat, 27 Dec 2014 07:54:50 -0800 (PST) alan wrote:
>
> >
> > On Saturday, December 27, 2014 9:12:20 AM UTC-5, Robert Heller wrote:
> >
> > > The article also seems to completely leave out the fact that MS-Word only runs under MS-Windows,
> >
> > This is simply not true: Word runs on the Mac OS too and has done for years.
> > It doesn't run under Linux though although the Web version
> > (http://www.dedoimedo.com/computers/office-online-linux.html) which is free,
> > in fact, does.
>
> There are some interoperatibilty issues between MacOS versions of MS-Word and
> the MS-Windows versions

Not sure how this addresses the basic point that Word is available for Mac too, and has pretty much always been. Interoperabilty isn't much of a problem with recent versions of Word in my experience anyway.

>
> Also, I don't believe it is normally possible to get MS-Word pre-installed on
> Macs, where most PC OEMs offer to pre-install MS-Office as part of the base
> install. For *many* people with Macs, MS-Office is more often not available
> than for people with MS-Windows PCs.

In any large institutional setting that provides Macs, this is certainly not true. All my colleagues and I use Macs as do many many others in my university and everyone gets Office installed on their machines as a matter of course. Most administrative documents are circulated in Word form, so it's hard to avoid it, even though I myself don't use it except for these purposes. All of my other work is done using LaTeX.

> To a certain extent, MS-Office implies
> MS-Windows.

My main point is just that this is basically not the case. Perhaps true for personal use computers, but given that the paper is addressing scientific research, it's unlikely to be true.

> And in the case of this article, only machines with MS-Windows
> were used (with MS-Office/MS-Word). There was no mention of MacOS. (There
> was mention of Linux, but I'm assuming that the people with Linux machines
> were using LaTeX, since no mention was made of OpenOffice or LibreOffice.)

Sure, but that was probably a function of the fact that German universities in my experience don't support Macs institutionally (although I'm sure there are exceptions.) It doesn't affect the basic results of the study (flawed though they are.)


Alan

[Joseph Wright]

On 27/12/2014 21:11, alan wrote:
>> There are some interoperatibilty issues between MacOS versions of MS-Word and
>> the MS-Windows versions
>
> Not sure how this addresses the basic point that Word is available for Mac too, and has pretty much always been. Interoperabilty isn't much of a problem with recent versions of Word in my experience anyway.

Try doing anything that uses DDE to store application data in a Word
file. For example, as a chemist I *require* ChemDraw figures. They can
be pasted into Word and stay editable *provided* I stick to either
Windows or Mac. Start a file on Windows, has to be kept on Windows,
start a file on Mac, has to be kept on Mac, otherwise you loose the
ability to edit the ChemDraws.
--
Joseph Wright

[Scott Pakin]

On 12/27/2014 11:46 AM, Torsten Bronger wrote:
>> I'm not sure as far as LaTeX is concerned, but TeX remains since
>> 1982 (more than 30 years) the leading typesetting system for
>> science, and will for the next decades for sure.
>
> Is there any solid evidence for this? I would be particular
> interested in statistics about how many journals accept which input
> formats.

It might make more sense to gather these statistics by publisher rather
than by journal. Some publishers produce a *lot* of journals -- and
conference proceedings; don't forget about those. One wouldn't want the
results biased by a few prolific publishers.

I believe most of the publishers who have published my papers request
both a PDF file and all source files (either LaTeX or Word) for final
journal submissions. I don't recall encountering a publisher that
allowed only one of LaTeX or Word but not the other. Hence, my guess
is that one could not find solid evidence for TeX or LaTeX being the
"leading typesetting system for science", but also that one probably
could not find solid evidence that Word serves that role, either.

Personally, I'm curious what typesetting systems the major scientific
publishers use internally. They all claim to use proprietary systems,
but I would guess that they're really using (La)TeX or Word plus a
bunch of in-house macros -- whether the copy editors realize that or
not.

-- Scott

[Torsten Bronger]

Hallöchen!

Scott Pakin writes:

> On 12/27/2014 11:46 AM, Torsten Bronger wrote:
>

> [...]

>
>> Is there any solid evidence for this? I would be particular
>> interested in statistics about how many journals accept which
>> input formats.
>
> It might make more sense to gather these statistics by publisher
> rather than by journal. Some publishers produce a *lot* of
> journals -- and conference proceedings; don't forget about those.
> One wouldn't want the results biased by a few prolific publishers.

I fail to see how gathering by publisher helps here.

Be that as it may: Once the data is there, you can analyse it in
various ways. Weighting with impact factor, number of
articles/pages per year etc.

If any publication investigating the best document format for
scientists is supposed to be significant, it must include such a
survey in my opinion.

[Joseph Wright]

On 28/12/2014 06:27, Scott Pakin wrote:
> Personally, I'm curious what typesetting systems the major scientific
> publishers use internally. They all claim to use proprietary systems,
> but I would guess that they're really using (La)TeX or Word plus a
> bunch of in-house macros -- whether the copy editors realize that or
> not.

ArborText is pretty popular. It's derived from a system which used TeX
as the 'back end' for typesetting but I believe they dropped TeX in
favour of their own proprietary system many years ago. Everything gets
converted to XML, in any case.
--
Joseph Wright

[Herbert Schulz]

Howdy,

I'll add a couple of comments.

While Word does offer sytle sheets I've virtually never seen anyone use
them. Even using a style sheet doesn't guarantee that the document will
have consistent formatting since there seems to be multiple, totally
independent ways to accomplish the same task and the style sheet
invariably misses some of these ways. In the end the document looks
terrible and has been hand-formatted along the way so the writer's
train of thought was interrupted.

Since LaTeX places it's journal formatting in a documentclass the
writer never interrupts the train of thought with hand formatting and
just uses abstract commands, e.g., \section{Section Title}, and lets
the documentclass take care of the formatting. Therefore writing LaTeX
(or other macro packages for TeX) documents is much MORE productive
than writing Word documents.

Finally, I've been able to take documents I've written in the late
1980s and early 1990s and produce a PDF or printed copy from them since
they are all simple ASCII text. I dare you to try the same with Word
where the file format has changes over the years at the manufacturer's
whim. Proprietary formats stink as far as historical documents are
concerned. Recent versions of Word seem to use a (pseudo-?)XML format
but not all of it's features are documented as far as I can tell.

All in all I'd suggest that people using LaTeX are MORE productive in
their writing than anyone using Word.

Good Luck,
Herb Schulz

[Axel Berger]

tsd wrote on Sat, 14-12-27 02:38:

>preventing researchers from producing documents in LaTeX would save time
>and money to maximize the benefit of research and development for both
>the research team and the public.

I've read the article now. The task was to reproduce the layout and
content of one or two professionally typeset pages. Assuming comparable
familiarity with both (which I don't have, so I'd swap extra work in
LaTeX for having to learn all new tricks) the point and click interface
is probably really the better one here. If you have exactly one and
only one headline of a given level, there is no onus on having to
format each single one individually.
The point is, once you have spent the time, can be up to a week, on
writing a class for all the niceties a given journal insists on, you
can then churn out articles by the dozen without the need to remember
any of those.
Teaching my literature database how to output a thebibliography
environment to exactly the specifications of the DAI took me several
days (on and off). Now, many years later, I have forgotten nearly all
those requirements and couldn't do it manually, but can print
references off by the hundred without a single mistake. Stuff like that
broken link in their own references simply could not happen.

A more realistic task would have been to start off with some pure ASCII
text and/or handwritten notes and requiring any pleasing and easily
legible result. With LaTeX and its defaults you're hard put and need to
put in a lot of effort to fail, with Office programs the reverse is
true. Yes, using them you CAN achieve very good results, but for those
the effort is more, not less than with with TeX.

Axel

[Torsten Bronger]

Hallöchen!

Axel Berger writes:

> [...]

>
> A more realistic task would have been to start off with some pure
> ASCII text and/or handwritten notes and requiring any pleasing and
> easily legible result.

FWIW, I don't need particularly pleasing reasults. They are
imported into the publisher's software nevertheless. To overstate
the case somewhat, I only need the letters to have the correct
ordering. Even the formulae only need to be correct, not pleasant.
That said, I don't write papers consisting of 40% equations.

I know that some publications are submitted camera-ready but I'm
pretty sure that that's a niche.

[Axel Berger]

Herbert Schulz wrote on Sun, 14-12-28 15:36:

>Recent versions of Word seem to use a (pseudo-?)XML format

What you get is an archive of about 20 files. In ODF the footnotes at
least come in the same file as the main text, in DOCX they are hidden
somewhere entirely different and their placement is anything but
obvious. In

http://berger-odenthal.de/upload/docx.zip (2.1 kB)

I have the same small text snippet from some DOCX I got as DOCX, as
ODF, as ASCII, and as LaTeX source. If that doesn't convince people,
nothing will.
Do not face towards the screen when viewing the DOCX unless your
stomach is completely empty - you have been warned.

>All in all I'd suggest that people using LaTeX are MORE productive in
>their writing than anyone using Word.

Probably true but not helpful. The test has been done and the results
published. They are probably not Stapels with invented data. So the
question is to find out just how their setup was rigged to produce the
intended result. I believe I gave one of the reasons a few posts back.

[Robert Heller]

Right. The article *IS* M$ FUD. The authors were just clever enough to hide
their connection to M$. Actually, they probably don't have a *real* connection
to M$, but instead have a 'connection in spirit' with M$ -- the giveaway was
the specificness of MS-Word rather then word processors in general, eg there
was no mention was made of Open Office or LibreOffice or of ODT format files.
They *effectively* started with the 'conclusion' that *specificly* MS-Word is
'easier and more productive', and then rigged the study to prove that
conclusion. I don't know if this was premeditated / deliberate or not. Plus
the *small* sample set tosses the statistical analysis out as meaningless.

The rigging givaway was that the people in the study were given a
*professionally typeset* copy to *reproduce*. Almost *all* *researchers*
(except maybe those with/getting a Typographical Design degree) are not
Typographical Designers. People using MS-Word might be *effectively*
*amature* Typographical Designers (this is the nature of how most people
actually use word processors). LaTeX users don't have to generally even
think about the Typographical Design of their documents. The authors of the
LaTeX macro package have already done the Typographical Design which LaTeX's
macros implement. Duh!

The *reality* is that a researcher who publishes for the set of journals for
his or her field will do a *one time* job of creating a documentclass (or more
likely modify an existing documentclass) to match each of those journals. The
'cost' (in time/effort) will be spread over many papers and actually the
documentclass / style files, might in fact be shared across many researchers
in the given field(s) for the given journal(s).

[Michael Shell]

On Sat, 27 Dec 2014 00:04:08 -0500 (EST)
Jack Ryan <nor...@remailer.cpunk.us> wrote:
> The most obvious oversight is that the article neglects the value of
> information acquired as a user advances.
>
> The struggles a LaTeX user faces are primarily the result of SUEs
> (stupid user errors), in which case they learn something of value that
> pays off in the long run.

Correct. After awhile, it is unreal how productive LaTeX can be after
the user develops the structures he/she uses the most. Unreal.

On Sun, 28 Dec 2014 08:36:04 -0600
Herbert Schulz <he...@wideopenwest.com> wrote:
> All in all I'd suggest that people using LaTeX are MORE productive in
> their writing than anyone using Word.

Ditto.

On Sat, 27 Dec 2014 07:12:22 -0600
Robert Heller <hel...@deepsoft.com> wrote:
> And I wonder how much M$ *paid* the researchers... I note that the article
> makes no mention of OpenOffice / LibrOfice, only *specifically* MS-Word.

Yep.



Cheers,

Mike Shell

[Axel Berger]

Torsten Bronger wrote on Sun, 14-12-28 23:25:

>They are imported into the publisher's software nevertheless.

I've admittedly never been in the position to have anything of mine
printed (since the school newsletter and student politics), but even
you will write lots of papers that are distributed and read as-is for a
smaller audience. I've had to read too many double spaced (in Microsoft
speak, which is really 240 % spaced) dissertations and it's an agony.

The same goes for simple stuff like handouts and using LaTeX I have for
several years typeset a local party newsletter with good results. The
chance I will read something I'm only partially interested in rises
steeply the less of a strain it is to read.

[Peter Wilson]

I'm hoping that there will be a response from Peter Flynn who had some
participation in the study.

Peter W.

[Peter Wilson]

What surprises me is that except for maths even the expert LaTeX users
were in almost all cases rated alongside or worse than novice Word users.

[Alain Ketterlin]

Peter Wilson <herrie...@earthlink.net> writes:

> What surprises me is that except for maths even the expert LaTeX users
> were in almost all cases rated alongside or worse than novice Word
> users.

LaTeX users made an average of 17 "typographic errors" for the
continuous text experiment. These errors include things like "header",
"columns"/"lines" (who knows what this means?), "spacing", "paragraph"
etc. Apparently, they were asked to reproduce exactly a single column of
text, plus header, and a single column title (at least for the
institution).

This experiment was obviously measuring the ability to format individual
elements, which is fairly easy on such a small scale with a
point-and-click interface, but takes quite a bit of time with LaTeX
(participants had 30 minutes). Ask the same people to type 15 pages of
text and give credit to consistency of layout, and the rating will
reverse.

The same goes for the table (a quarter page of text on top, with the
table at the bottom, with huge space in between). The table itself has
lots of merged cells. The equations were numbered and left aligned (they
probably required the numbers to be (62) etc.).

Overall the experiment was all about visual formatting, designed by
people who think this is the most important aspect of technical writing.
Basically, the question they asked themselves was: can you do with LaTeX
the kind of things you can do with Word? (Even though this has not much
to do with technical writing, and probably is not even close to what
serious users of Word do.)

That's why the whole experiment doesn't make any sense.

-- Alain.

[Torsten Bronger]

Hallöchen!

Peter Wilson writes:

> What surprises me is that except for maths even the expert LaTeX
> users were in almost all cases rated alongside or worse than
> novice Word users.

While I don't know the details of the investigation, the keyboard
math syntax of current Word versions is simpler than LaTeX's for
most use cases.

[Robert Heller]

At Mon, 29 Dec 2014 18:37:38 +0000 Peter Wilson <herrie...@earthlink.net> wrote:

>
> What surprises me is that except for maths even the expert LaTeX users
> were in almost all cases rated alongside or worse than novice Word users.

Given the *nature* of the test, this is not actually suprising. This is how
the test was 'rigged'. Rather than use a 'standard' documentclass to generate
good quality output, the people in the study were told to reproduce a
professionally typeset document. Most mess-word users do 'visual design',
that is that are using mess-word to directly produce a specific visual look.
With LaTeX, one uses a pre-defined documentclass. Using LaTeX in a 'visual
design' mode is going to be tough for all but uber-expert-guru type LaTeX
users. You will end up using lots of random TeX primitives or defining lots
of commands and macros and probably fumbling with lots of low-level LaTeX
macro programming, which *most* researchers don't need to do most of the time.
Since mess-word users commonly are 'messing with' random direct formatting
stuff (using all the nonsense toolbar crapola) and getting instant visual
feedback as they do that, it makes perfect sense that the mess-word users were
able to duplicate the typeset document easily.

[GL]

Adobe InDesign is a very good software as far as i know.

You may have a look at this:
https://cel.archives-ouvertes.fr/cel-00519301v2/document

(it is an lesson in "analyse" - in french, but the formatting,
in LaTeX, is really professional !)

Sincerely.

[Axel Berger]

Joseph Wright wrote on Sun, 14-12-28 09:44:
>ArborText is pretty popular.

Well, it is responsible for the broken link in the article we're
talking about -- something that jsut could't happen with LaTeX.

[paulschlesinger]

On Friday, December 26, 2014 7:39:01 PM UTC-6, tsd wrote:
> "Given these numbers it remains an open question to determine the amount of taxpayer money that is spent worldwide for researchers to use LaTeX over a more efficient document preparation system, which would free up their time to advance their respective field. Some publishers may save a significant amount of money by requesting or allowing LaTeX submissions because a well-formed LaTeX document complying with a well-designed class file (template) is much easier to bring into their publication workflow. However, this is at the expense of the researchers' labor time and effort. We therefore suggest that leading scientific journals should consider accepting submissions in LaTeX only if this is justified by the level of mathematics presented in the paper. In all other cases, we think that scholarly journals should request authors to submit their documents in Word or PDF format. We believe that this would be a good policy for two reasons. First, we think that the appearance of the text is secondary to the scientific merit of an article and its impact to the field. And, second, preventing researchers from producing documents in LaTeX would save time and money to maximize the benefit of research and development for both the research team and the public."
>
> http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0115069

I am surprised that the authors had the time for this study but I am not surprised it was published as the stated PLOS editorial imperative is not "is the article significant."
The authors appear not to have included the cost of purchasing the upgrades to Word in their analysis and the time spent learning the changes for the user. I have manuscripts and lectures done in 1979-80 that still compile without change, although with fewer embellishments. I stopped using Word for production work in 1979 when a "blue screen of death" destroyed a day of editing of a lecture outline. As I understand the program design Word is a page editor, which explains a lot of the difficulties in getting formatting and Figures to behave. However I do not regularly use Word except when editing with collaborators and then I usually convert documents. I find this less time consuming that trying to determine the current keys for for super/subscripts or reformatting a paragraph. I also edit in emacs and vim because the terminal 8 point fonts are much better on my old eyes than Words version of wysiwyg. Also these editors seem to understand how to scroll precisely. As a page editor Word is excellent subject to the above caveats but since I no longer make one page flyers for my department and do need to review my prior manuscripts without finding an old version of software and have never lost a document other than in a proprietary format, I will continue to use TeX/LaTeX where I choose to employ it. I really do not give a damn what journals require, documents can be converted.
PHS

[Peter Wilson]

On 29/12/14 21:04, Robert Heller wrote:
> At Mon, 29 Dec 2014 18:37:38 +0000 Peter Wilson <herrie...@earthlink.net> wrote:
>
>>
>> What surprises me is that except for maths even the expert LaTeX users
>> were in almost all cases rated alongside or worse than novice Word users.
>
> Given the *nature* of the test, this is not actually suprising. This is how
> the test was 'rigged'. Rather than use a 'standard' documentclass to generate
> good quality output, the people in the study were told to reproduce a
> professionally typeset document. Most mess-word users do 'visual design',
> that is that are using mess-word to directly produce a specific visual look.
> With LaTeX, one uses a pre-defined documentclass. Using LaTeX in a 'visual
> design' mode is going to be tough for all but uber-expert-guru type LaTeX
> users. You will end up using lots of random TeX primitives or defining lots
> of commands and macros and probably fumbling with lots of low-level LaTeX
> macro programming, which *most* researchers don't need to do most of the time.
> Since mess-word users commonly are 'messing with' random direct formatting
> stuff (using all the nonsense toolbar crapola) and getting instant visual
> feedback as they do that, it makes perfect sense that the mess-word users were
> able to duplicate the typeset document easily.
>
>>

I have now looked more closely at the continuous text sheet that had to
be reproduced. I haven't tried to reproduce it myself but although being
an `expert' LaTeX user I would expect to fail miserably. First off, I'm
a two fingered typist who tends to type typos. Then, the text layout
obviously doesn't match any of the common LaTeX classes so I would have
to spend some time in deciding just what LaTeX constructs I would use
(or create). It appears to be a two column article, except that the
testers have eliminated the second column. Thus in LaTeX one would go
for a two-column layout but then the title, which spans more than a
column, has to be catered for. Further, although I printed out the sheet
on A4 paper the single column is wider than half the width of the paper;
how is this to be accounted for? And so on. I could well imagine the 30
minutes being up before I had produced any words.

I have managed to avoid MS ever since it appeared so have no idea how
well the MS users would have coped with the challenge. I have the
feeling, though, that once the LaTeX user had got everything set up
producing further sheets in the same style would essentially come down
to just typing the text.

Peter W.

[Robert Heller]

Right. A researcher writting real papers for real journals would use a class
or style file customized for each journal. This class or style file would
probably either be supplied by the journal(s) or jointly created by
researchers writting for the journal(s) in question. Thus this design work
would be spread across many papers, reducing the cost per paper. The error
rate and the time spent *actually writing the paper* would also go down. With
MS-Word, it is 'normal' for people to spend lots of time 'fiddling' with the
formatting (this is the nature of the beast). 'Expert' MS-Word uses become
'expert' at using the many formatting 'tools'. LaTeX users generally don't
spend much time or effort fiddling with formatting, since the *formatting*
(really typographical design) is implemented by class and style files. The
study the article describes on many levels is really a meaningless comparison
and was effectively 'rigged' in favor of MS-Word.

>
> I have managed to avoid MS ever since it appeared so have no idea how
> well the MS users would have coped with the challenge. I have the
> feeling, though, that once the LaTeX user had got everything set up
> producing further sheets in the same style would essentially come down
> to just typing the text.

Word Processors are (effectively) all about visual design with near instant
graphical feedback as the user 'fiddles' with formatting choises (provided by
elaborate toolbars). One can select fonts, font sizes, bold, italics, etc.
Once can select a chuck of text and move the text around, make the columns
wider or narrower, etc. In many ways, word processors are NOT about the
*words*, but about the graphical representation.

>
> Peter W.

[Scott Pakin]

On 12/28/2014 01:44 AM, Joseph Wright wrote:
> ArborText is pretty popular. It's derived from a system which used TeX
> as the 'back end' for typesetting but I believe they dropped TeX in
> favour of their own proprietary system many years ago. Everything gets
> converted to XML, in any case.

Interesting. I recall once receiving some proofs from a publisher
that certainly looked like TeX output. I sent the copy editor some
LaTeX code to replace a piece of flawed typesetting I spotted,
figuring that was the easiest path to fixing the typesetting. The
copy editor replied that the document was not being typeset with LaTeX
but rather with a proprietary system. Maybe it was ArborText back
when it used TeX as the back end.

-- Scott

[Torsten Bronger]

Hallöchen!

Robert Heller writes:

> [...]
>
> [...] LaTeX users generally don't spend much time or effort

> fiddling with formatting, since the *formatting* (really
> typographical design) is implemented by class and style

> files. [...]

My observations have been quite different. As the local LaTeX
wizard, I have seen many LaTeX documents of otherwise very
intelligent collegues. All of their preambles were monstrous, and
none of them was correct. There must have been a lot of frustrating
hours being wasted on these preambles for sure.

[Robin Fairbairns]

Peter Wilson <herrie...@earthlink.net> writes:

> What surprises me is that except for maths even the expert LaTeX users
> were in almost all cases rated alongside or worse than novice Word
> users.

the expert latex users were presumably writing what the "study"'s
authors asked for, and were showing up the flaws in the spec.

presumably the flaws were part of a design process to make the output
look like word output. word output is typically crappy, and the study
shows that it's difficult to make latex output as bad as that.
--
Robin Fairbairns, Cambridge

[Dr Eberhard Lisse]

Some arguments about useability, such as spelling errors and such make
sense.

But LyX is an answer.

el



On 2014-12-27 03:38 , tsd wrote:
[...]
> http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0115069


--
if you want to reply, replace nospam with my initials

[Nicolas Richard]

Torsten Bronger <bro...@physik.rwth-aachen.de> writes:
> As the local LaTeX
> wizard, I have seen many LaTeX documents of otherwise very
> intelligent collegues. All of their preambles were monstrous, and
> none of them was correct. There must have been a lot of frustrating
> hours being wasted on these preambles for sure.

I agree on the facts, which I observed also, but the conclusion I draw
is different : I think the monstrous preambles are built over the years
by multiple generation of people, each adding or modifying a few lines
that they found somewhere. I don't think they actually spend a lot of
time writing their --usually self-contradictory, by the way-- preambles.

I suggest we conduct a study on this matter :)

--
Nicolas Richard

[Peter Flynn]

On 12/27/2014 01:38 AM, tsd wrote:
> "Given these numbers it remains an open question to determine the
> amount of taxpayer money that is spent worldwide for researchers to
> use LaTeX over a more efficient document preparation system, which
> would free up their time to advance their respective field.

The key here is the definition of "more efficient". Efficiency is
normally regarded as a measure of how well the effort achieves the goal;
that is, "how well" in this sense is a qualitative measure (there are
specific quantitative measures in its more restricted sense in
engineering, physics, etc where accepted formulae exist).

The use of LaTeX by poorly-instructed users, or by users with
insufficient understanding of the concept of a structured document, will
typically be less efficient (ie take more effort) than the use of Word
by the same users, no matter how poorly instructed, because the goal of
these users is to create a visually attractive or conformant document
(in their eyes). The goal of the well-instructed user with a proper
understanding of structure is to create a robust and durable document
that will outlast all other formats as well as be visually attractive or
conformant.

This paper only addresses the problems faced by the relatively novice
user with narrowly-defined goals, for which keystroke-level modelling is
useful. It would require a completely different methodology to address
the wider applicability of LaTeX vs Word by more experienced users,
which would include non-interactive factors such as document
persistence, mechanical reliability, and content re-use.

> "Some publishers may save a significant amount of money by requesting
> or allowing LaTeX submissions because a well-formed LaTeX document
> complying with a well-designed class file (template) is much easier
> to bring into their publication workflow. However, this is at the
> expense of the researchers' labor time and effort.

This is a good illustration of the case in point: among novice users,
there is significant time and effort required because there is a dual
learning curve: (a) for the concepts of structured documents and the
automation that is possible in LaTeX; and (b) the set of LaTeX commands
required to implement (a). This is not the case with seasoned authors
(Word or LaTeX) equipped with the proper tools, and this would require a
different approach.

> "We therefore suggest that leading scientific journals should
> consider accepting submissions in LaTeX only if this is justified by
> the level of mathematics presented in the paper. In all other cases,
> we think that scholarly journals should request authors to submit
> their documents in Word or PDF format.

PDF is unlikely to gain acceptance from publishers because entry to the
workflow at the PDF stage presupposes that the document has been
competently formatted by a typesetter. While it is perfectly possible
for the seasoned user of LaTeX to achieve this, it may require even the
seasoned Word user additional effort.

> "We believe that this would be a good policy for two reasons. First,
> we think that the appearance of the text is secondary to the
> scientific merit of an article and its impact to the field.

There may be many Word and LaTeX users who disagree with this. If the
scientific merit cannot be ascertained because it is obscured by
incompetent formatting, it may fail this test, but that is not a reason
for discarding the test, it's an argument for making better systems.

> "And, second, preventing researchers from producing documents in
> LaTeX would save time and money to maximize the benefit of research
> and development for both the research team and the public."

The amount of time and money saved may be inconsequential compared with
the overall cost of a research project, especially in view of the
subsequent accessibility risk for documents which lack reusable markup
(both LaTeX and Word).

///Peter
--
Claimer: I was a reviewer of this paper.
I will address some of the other comments in separate messages.

[Torsten Bronger]

Hallöchen!

Nicolas Richard writes:

> Torsten Bronger <bro...@physik.rwth-aachen.de> writes:
>
>> As the local LaTeX wizard, I have seen many LaTeX documents of
>> otherwise very intelligent collegues. All of their preambles
>> were monstrous, and none of them was correct. There must have
>> been a lot of frustrating hours being wasted on these preambles
>> for sure.
>
> I agree on the facts, which I observed also, but the conclusion I
> draw is different : I think the monstrous preambles are built over
> the years by multiple generation of people, each adding or
> modifying a few lines that they found somewhere.

This is true. But it doesn't matter how those preambles came into
existence. They are a burden anyway. The students struggle with
these preambles before finally asking me for help.

In one case, a PhD student eventually accepted that his document was
errorneous, and pressed "r <RET>" every time he updated the DVI/PDF.
It didn't affect the output.

> [...]

>
> I suggest we conduct a study on this matter :)

It would be an interesting survey, indeed. I also doubt that LaTeX
users fiddle less with the local layout than Word users. They just
fiddle in a different way I suspect. For example, the placement of
floats is a constant source of amusement for LaTeX users. Tables
can be much trickier than in WYSIWYG systems. All of this costs
time and patience.

I think especially those users who try to treat LaTeX like a word
processor ("I want to have this and that this way, dammit!") waste
time. It may be their fault but this doesn't help either.

[Axel Berger]

Torsten Bronger wrote on Wed, 14-12-31 13:10:

>Tables can be much trickier than in WYSIWYG systems.

I've never seen that. Take a CSV, replace ; by &, done. As with girls,
those who are beutiful in the first place need very little makeup on
top.

>("I want to have this and that this way, dammit!")

I plead guilty to that. It sometimes takes a ridiculous amount of time.
But then you have done it just once, can forget all about it, and every
single document will look perfect every time.

[Robert Heller]

At Wed, 31 Dec 2014 15:26:00 +0100 Axel_...@b.maus.de (Axel Berger) wrote:

>
> Torsten Bronger wrote on Wed, 14-12-31 13:10:
> >Tables can be much trickier than in WYSIWYG systems.
>
> I've never seen that. Take a CSV, replace ; by &, done. As with girls,
> those who are beutiful in the first place need very little makeup on
> top.

I guess from an 'instant gratification' POV, a table in LaTeX would probably
look 'ugly'. And *sparce* tables can be tricky (knowing when and where to put
the extra '&'s).

>
> >("I want to have this and that this way, dammit!")
>
> I plead guilty to that. It sometimes takes a ridiculous amount of time.
> But then you have done it just once, can forget all about it, and every
> single document will look perfect every time.

One other issue relating to mess-word vs. LaTeX: mess-word will let you do
things that are really bad from a typographical POV, but LaTeX won't let you
do such things (at least not without major convolutions). One example: if you
have an document in 2-column mode and want to put in an equation or figure
that spans both columns. This is typographicly bad since when the figure or
equation spans the two columns there is not a clear one-on-one relationship
between the figure or equation and specific piece of the text. LaTeX *forces*
such equations or figures onto pages with only such equations or figures.

[Michael Shell]

On Wed, 31 Dec 2014 11:52:27 +0100
"Nicolas Richard" <theonewith...@yahoo.fr> wrote:

> I think the monstrous preambles are built over the years
> by multiple generation of people, each adding or modifying a few lines
> that they found somewhere. I don't think they actually spend a lot of
> time writing their --usually self-contradictory, by the way-- preambles.

FWIW, that's been my observation too - I've seen authors that loaded
everything but the kitchen sink, but only one or two of those packages
were actually used.

When asked why they loaded all those packages, they are usually unable
to provide an answer and are often surprised that I would ask such
a strange question - it was simply that way when they got their starter
files. They think nothing of it.

If the average LaTeX user knew what went on behind the scenes to
get all those packages to work together, they would be a little
more hesistant to go hog wild loading things. Really, I am amazed
that most all of the modern LaTeX packages are compatible with
each other.

Maybe there is a need to better educate newbies about the "preamble
bloat" issue.


Cheers,

Mike Shell

[Rusi Mody]

On Thursday, January 1, 2015 2:32:44 AM UTC+5:30, Robert Heller wrote:

> At Wed, 31 Dec 2014 15:26:00 +0100 Axel Berger wrote:
>
> >
> > Torsten Bronger wrote on Wed, 14-12-31 13:10:
> > >Tables can be much trickier than in WYSIWYG systems.
> >
> > I've never seen that. Take a CSV, replace ; by &, done. As with girls,
> > those who are beutiful in the first place need very little makeup on
> > top.
>
> I guess from an 'instant gratification' POV, a table in LaTeX would probably
> look 'ugly'. And *sparce* tables can be tricky (knowing when and where to put
> the extra '&'s).

People editing (la)tex with emacs can have the cake¹ and eat it too.

Basically emacs-org-mode has spreadsheet behavior.
One can use that to create the table.
And then 'compile' it into a latex table (within the same file)

Here is a solution I had given to the python list:
https://groups.google.com/d/topic/comp.lang.python/dV84Id4E2kM/discussion
I can try producing a similar one for latex if that's desired and not too OT

--------------
¹ Speaking of cake... Happy new year

[Axel Berger]

Robert Heller wrote on Wed, 14-12-31 22:02:

>And *sparce* tables can be tricky (knowing when and where to put the
>extra '&'s).

If you think not about the table being sparse but the fields being
empty it becomes obvious and your CSV with the data already has all
those multiple delimiters in the right place.

>LaTeX *forces* such equations or figures onto pages with only such
>equations or figures.

No, in twocolumn mode only to the top or bottom of a page. That goes
for images, for equations there is nothing equivalent but you can place
anything in a table* environment as long as you don't need a caption.

[Khaled Hosny]

On Sunday, December 28, 2014 8:27:13 AM UTC+2, Scott Pakin wrote:
> On 12/27/2014 11:46 AM, Torsten Bronger wrote:
> >> I'm not sure as far as LaTeX is concerned, but TeX remains since
> >> 1982 (more than 30 years) the leading typesetting system for
> >> science, and will for the next decades for sure.
> >
> > Is there any solid evidence for this? I would be particular
> > interested in statistics about how many journals accept which input
> > formats.
>
> It might make more sense to gather these statistics by publisher rather
> than by journal. Some publishers produce a *lot* of journals -- and
> conference proceedings; don't forget about those. One wouldn't want the
> results biased by a few prolific publishers.
>
> I believe most of the publishers who have published my papers request
> both a PDF file and all source files (either LaTeX or Word) for final
> journal submissions. I don't recall encountering a publisher that
> allowed only one of LaTeX or Word but not the other. Hence, my guess
> is that one could not find solid evidence for TeX or LaTeX being the
> "leading typesetting system for science", but also that one probably
> could not find solid evidence that Word serves that role, either.

>
> Personally, I'm curious what typesetting systems the major scientific
> publishers use internally. They all claim to use proprietary systems,
> but I would guess that they're really using (La)TeX or Word plus a
> bunch of in-house macros -- whether the copy editors realize that or
> not.

I know a major scientific publisher that does not even accept LaTeX files from authors, yet the printed version is eventually typeset with LaTeX.

[Phillip Helbig (undress to reply)]

In article <9bff8d2b-6bd8-4924...@googlegroups.com>,

Khaled Hosny <dr.khal...@gmail.com> writes:

> I know a major scientific publisher that does not even accept LaTeX
> files from authors, yet the printed version is eventually typeset with
> LaTeX.

The reverse exists as well: submission in LaTeX, then final typesetting
in something else. Some do both, some do neither.

Note: When this text becomes available on the world-wide web,
"submission in LaTeX" will attract a surprising number of referals from
search engines. :-)

[Torsten Bronger]

Hallöchen!

Phillip Helbig (undress to reply) writes:

> [...]

>
> Note: When this text becomes available on the world-wide web,
> "submission in LaTeX" will attract a surprising number of referals
> from search engines. :-)

Because it's so rarely found elsewhere? :-P

[Athel Cornish-Bowden]

On 2015-01-01 21:16:19 +0000, Phillip Helbig (undress to reply said:

> [ … ]

> Note: When this text becomes available on the world-wide web,
> "submission in LaTeX" will attract a surprising number of referals from
> search engines. :-)

When I read this I thought you were referring to pages like this:
http://www.youjizz.com/videos/submission-in-latex-171645.html, but the
first two pages from Google were all about journal submission.


--
athel

[John Harper]

Torsten Bronger wrote:

> Hallöchen!
>
> Robert Heller writes:
>
>> [...]
>>
>> [...] LaTeX users generally don't spend much time or effort
>> fiddling with formatting, since the *formatting* (really
>> typographical design) is implemented by class and style
>> files. [...]
>
> My observations have been quite different. As the local LaTeX
> wizard, I have seen many LaTeX documents of otherwise very
> intelligent collegues. All of their preambles were monstrous, and
> none of them was correct. There must have been a lot of frustrating
> hours being wasted on these preambles for sure.

Those observations may be untypical: some of Torsten's colleagues may have
perfectly good preambles and don't need to ask him for help.

I have preamble trouble myself at present: last Friday I sent a book review
to Mathematical Reviews. The PDF from my LaTeX software looked OK, but
theirs did not. Usually theirs works OK for me. I have asked their LaTeX
wizards for help, but it's still the weekend in USA so I'm not asking this
newsgroup (yet). Of course the Math. Revs. wizards may well tell me the same
things that Torsten tells his colleagues :-)

--
John Harper

[Marc van Dongen]

On Wednesday, 31 December 2014 21:02:44 UTC, Robert Heller wrote:
> At Wed, 31 Dec 2014 15:26:00 +0100 Axel_...@b.maus.de (Axel Berger) wrote:
>
> >
> > Torsten Bronger wrote on Wed, 14-12-31 13:10:
> > >Tables can be much trickier than in WYSIWYG systems.
> >
> > I've never seen that. Take a CSV, replace ; by &, done. As with girls,
> > those who are beutiful in the first place need very little makeup on
> > top.
>
> I guess from an 'instant gratification' POV, a table in LaTeX would probably
> look 'ugly'. And *sparce* tables can be tricky (knowing when and where to put
> the extra '&'s).
>
> >
> > >("I want to have this and that this way, dammit!")
> >
> > I plead guilty to that. It sometimes takes a ridiculous amount of time.
> > But then you have done it just once, can forget all about it, and every
> > single document will look perfect every time.
>
> One other issue relating to mess-word vs. LaTeX: mess-word will let you do
> things that are really bad from a typographical POV, but LaTeX won't let you
> do such things (at least not without major convolutions). One example: if you
> have an document in 2-column mode and want to put in an equation or figure
> that spans both columns. This is typographicly bad since when the figure or
> equation spans the two columns there is not a clear one-on-one relationship
> between the figure or equation and specific piece of the text. LaTeX *forces*
> such equations or figures onto pages with only such equations or figures.

I had a quick look at the paper but I don't see the point in investigating a LaTeX user's ability to reproduce some numbered displayed equations with labels starting at 60. Most users won't know how to do this and that's a good thing. I don't see why it's relevant/important a LaTeX user is worse at this than an X user.

Similarly, (others also mentioned this) expecting that a LaTeX user is capable of mimicing a page layout is not realistic and I cannot see any value in testing this ability or drawing any conclusions from it (other than that they're statistically worse at it).

What is nice about LaTeX is that the markup lets users format their text/equations quickly.

What's even nicer is that they can automate frequently occurring tasks by implementing user-defined macros; developing macros takes time but once a good macro is available, it can be used over and over again. Using macros also is good when it comes to developing a document and maintaining it.

I could provide more comments/criticism but I don't have the time...

Just my opinion.

Regards,


Marc van Dongen

[Robert Heller]

At Mon, 5 Jan 2015 04:20:10 -0800 (PST) Marc van Dongen <don...@cs.ucc.ie> wrote:

>
> On Wednesday, 31 December 2014 21:02:44 UTC, Robert Heller wrote:

> > At Wed, 31 Dec 2014 15:26:00 +0100 Axel_...@b.maus.de (Axel Berger) wr=
> ote:
> >=20
> > >=20

> > > Torsten Bronger wrote on Wed, 14-12-31 13:10:
> > > >Tables can be much trickier than in WYSIWYG systems.

> > >=20
> > > I've never seen that. Take a CSV, replace ; by &, done. As with girls,=
> =20
> > > those who are beutiful in the first place need very little makeup on=20
> > > top.
> >=20
> > I guess from an 'instant gratification' POV, a table in LaTeX would proba=
> bly=20
> > look 'ugly'. And *sparce* tables can be tricky (knowing when and where t=
> o put=20
> > the extra '&'s).
> >=20
> > >=20

> > > >("I want to have this and that this way, dammit!")

> > >=20
> > > I plead guilty to that. It sometimes takes a ridiculous amount of time.=
> =20
> > > But then you have done it just once, can forget all about it, and every=
> =20

> > > single document will look perfect every time.

> >=20
> > One other issue relating to mess-word vs. LaTeX: mess-word will let you d=
> o=20
> > things that are really bad from a typographical POV, but LaTeX won't let =
> you=20
> > do such things (at least not without major convolutions). One example: i=
> f you=20
> > have an document in 2-column mode and want to put in an equation or figur=
> e=20
> > that spans both columns. This is typographicly bad since when the figure=
> or=20
> > equation spans the two columns there is not a clear one-on-one relationsh=
> ip=20
> > between the figure or equation and specific piece of the text. LaTeX *fo=
> rces*=20

> > such equations or figures onto pages with only such equations or figures.
>

> I had a quick look at the paper but I don't see the point in investigating =
> a LaTeX user's ability to reproduce some numbered displayed equations with =
> labels starting at 60. Most users won't know how to do this and that's a go=
> od thing. I don't see why it's relevant/important a LaTeX user is worse at =

> this than an X user.
>

> Similarly, (others also mentioned this) expecting that a LaTeX user is capa=
> ble of mimicing a page layout is not realistic and I cannot see any value i=
> n testing this ability or drawing any conclusions from it (other than that =

> they're statistically worse at it).

Right. The paper pretty obviously played all of the *same* sorts of 'games'
early Linux vs. MS-Windows FUD studies played. The study was *intentionally*
rigged to make MS-Word 'look better', using a non-real world case. It was an
Apples vs. Oranges comparison where the goal was to test for the more orange
fruit.

>
> What is nice about LaTeX is that the markup lets users format their text/eq=
> uations quickly.
>
> What's even nicer is that they can automate frequently occurring tasks by i=
> mplementing user-defined macros; developing macros takes time but once a go=
> od macro is available, it can be used over and over again. Using macros als=
> o is good when it comes to developing a document and maintaining it.=20

>
> I could provide more comments/criticism but I don't have the time...
>
> Just my opinion.
>
> Regards,
>
>
> Marc van Dongen
>

[Marc van Dongen]

On Monday, 5 January 2015 13:19:03 UTC, Robert Heller wrote:

> Right. The paper pretty obviously played all of the *same* sorts of 'games'
> early Linux vs. MS-Windows FUD studies played. The study was *intentionally*
> rigged to make MS-Word 'look better', using a non-real world case. It was an
> Apples vs. Oranges comparison where the goal was to test for the more orange
> fruit.

I forgot to mention one thing. The displayed equations in the tests in the paper don't have any end-of-sentence punctuation symbols, which they _should_ have. It clearly demonstrates the paper is about being good at copying/reproducing/typesetting a given piece of text, not about doing it right!

Regards,


Marc van Dongen

[John Harper]

Some good authors (or editors?) disagree with you. Two Cambridge University
Press books on fluid mechanics have: punctuation after almost all displayed
equations (Drazin & Reid), and after almost none (Pozrikidis).

--
John Harper

[Luis Rivera]

On Friday, 26 December 2014 19:39:01 UTC-6, tsd wrote:
> "Given these numbers it remains an open question to determine the amount of taxpayer money that is spent worldwide for researchers to use LaTeX over a more efficient document preparation system, which would free up their time to advance their respective field. [...] First, we think that the appearance of the text is secondary to the scientific merit of an article and its impact to the field. And, second, preventing researchers from producing documents in LaTeX would save time and money to maximize the benefit of research and development for both the research team and the public."
>

Enrico Gregorio has shown in TUGBoat

https://www.tug.org/TUGboat/tb26-3/tb84gregorio.pdf

that many LaTeX users [mis|ab]use LaTeX precisely because they have too much access to the typesetter [TeX], and thus too much control of the final output; so they unconsciously try to use LaTeX as W*rd (or W*rd as LaTeX, from the other side).

I dare to say that the main flaw in the experiment reported in the paper is precisely that the subjects are asked to act as book designers, simply using different software to imitate a particular output: so the discussion quickly turns into which software produces the best output easier. However, both approaches (LaTeX vs *Office) miss the point altogether, already mentioned, though incidentally, in the report: the point being that content producers should not worry too much about the aspect of the document: it will be changed eventually by the publishers anyway; if they do worry, they are wasting their time (and taxpayer money, incidentally) trying to achieve a particular effect, either online or in paper, regardless of whether they use ConTeXt or Scribus. They are doing the publisher's job.

The alternative proposed somewhere else in this thread is XML. That is wrong too, imnsho: for XML is designed for machine and not human consumption. If you ask me, the least intrusive approach for an author is the use of something like WikiMarkup: a whole Encyclopedia is being edited by the second, as we speak, on this tiny markup. Leave the details of page layout or hyphenation to the publishers; whether they use HTML5, Plain TeX or InDesign shouldn't matter.

My penny thoughts on this issue,

Luis.

[Rusi Mody]

On Tuesday, January 6, 2015 7:36:56 AM UTC+5:30, Luis Rivera wrote:
> The alternative proposed somewhere else in this thread is XML. That is wrong too, imnsho: for XML is designed for machine and not human consumption. If you ask me, the least intrusive approach for an author is the use of something like WikiMarkup: a whole Encyclopedia is being edited by the second, as we speak, on this tiny markup. Leave the details of page layout or hyphenation to the publishers; whether they use HTML5, Plain TeX or InDesign shouldn't matter.

+1 from me -- with the following caveat.
Here is the wikipedia page on (ε,_δ)-definition_of_limit
https://en.wikipedia.org/wiki/%28%CE%B5,_%CE%B4%29-definition_of_limit

Editing it produces this excerpt

(\forall \varepsilon > 0)(\exists \ \delta > 0) (\forall x \in D)(0 < |x - c | < \delta \ \Rightarrow \ |f(x) - L| < \varepsilon)

Now compare it with

∀ ε > 0 (∃ δ > 0) (∀ x ∈ D) (0 < |x - c| < δ ⇒ |f(x) - L| < ε

In this age of unicode when we have xetex/luatex why do we use the first?

Hopefully most people would agree the latter is more readable than the former.
The questions that remain are

1 Typing it in.
2 Is it close to luatex/xetex?

For 1., Ive just recently discovered
https://github.com/rrthomas/pointless-xcompose/

For 2. I'd welcome help/suggestions ;-)
For 2 I would be happy for help ;-)

[Axel Berger]

Luis Rivera wrote on Tue, 15-01-06 03:06:

>the point being that content producers should not worry too much about
>the aspect of the document: it will be changed eventually by the
>publishers anyway;

That's only partially true. I frequently get unpublished drafts to read
and while those need not be perfect in all minutiae, they often are so
ugly as to make reading them notably hard. LaTeX makes it hard to make
presentations and handouts as bad as other tools often do.

And publishers do not ask for Word. What I really find is something
like:
"Manuscripts are accepted in contemporary standard formats (*.rtf,
*.odt, *.doc(x), *.wpd), ...
Text has to be unformatted, i.e. no justified text, no hyphenation, no
indentation ..." [DGUF]

To me that's no .odt or whatever but pure and simple ASCII.txt, but
with footnotes tied to their markers. You can approximate it in a
Word.doc, but in principle it's much nearer mininmally marked LaTeX
source or markdown.
There are so many people, for whom MS-Word is the only editor they know
and have ever used, putting it in those terms probabaly saves them a
lot support effort, and that's the only reason they do it

Axel

[Axel Berger]

Rusi Mody wrote on Tue, 15-01-06 03:24:

>(\forall \varepsilon > 0)(\exists \ \delta > 0) (\forall x \in D)(0 <
>|x - c | < \delta \ \Rightarrow \ |f(x) - L| < \varepsilon)
>
>Now compare it with
>

>Gammيا ?OE > 0 (Gammيق ?oe > 0) (Gammيا x Gammيي D) (0 < |x - c| < ?oe Gammهئ |f(x) - L| <
>?OE

>
>In this age of unicode when we have xetex/luatex why do we use the first?

Does that answer your question?

Axel

[Julian Bradfield]

On 2015-01-06, Rusi Mody <rusto...@gmail.com> wrote:
> (\forall \varepsilon > 0)(\exists \ \delta > 0) (\forall x \in D)(0 < |x - c | < \delta \ \Rightarrow \ |f(x) - L| < \varepsilon)
>
> Now compare it with
>
> ∀ ε > 0 (∃ δ > 0) (∀ x ∈ D) (0 < |x - c| < δ ⇒ |f(x) - L| < ε
>
> In this age of unicode when we have xetex/luatex why do we use the first?

Well over a decade ago, I went to a considerable amount of trouble
(building on work of Renaud Marlet) to build a mode for (X)Emacs that
let me see the latter (with simple compose-key inputs) instead of the
former.
I used it for a couple of years, and then gave up (though some of the
work found its way into the x-symbol package, which I believe is still
around).

Why? It's just not worth the effort. If you speak LaTeX, it's hardly
any easier to read the latter than the former; and when you really
want to read it for content, you read the typeset output, whether on
screen or on paper.

Where I *do* use Unicode is in phonology/phonetics - it's easier to
see phonetic letters as such, because the names tend to be long.

Though even there, I use it moderately. Here's a line of text from a
paper:

the click ‹ǃ̬qχ’› can be analysed as /(ǃ⊗ᴋ̬qχ’)/.

This is quite readable, right? Nonetheless, the source says

the click \Phon{\VcdEjcFrc{ǃ}} can be analysed as \Phon{/\concc{ǃ}{\Vcd{\ac}\EjcFrc{}}/}.

Why? Because the paper discusses, amongst other things, several
different formulations and notations; and my chances of typing the
wrong series of diacritical marks are much higher than my chances of
mistyping a descriptive control sequence such as \VcdEjcFrc (voiced
ejective fricate); and my notation for \concc changed a few times.

[Rusi Mody]

On Tuesday, January 6, 2015 8:30:05 PM UTC+5:30, Julian Bradfield wrote:

> On 2015-01-06, Rusi Mody wrote:
> > (\forall \varepsilon > 0)(\exists \ \delta > 0) (\forall x \in D)(0 < |x - c | < \delta \ \Rightarrow \ |f(x) - L| < \varepsilon)
> >
> > Now compare it with
> >
> > ∀ ε > 0 (∃ δ > 0) (∀ x ∈ D) (0 < |x - c| < δ ⇒ |f(x) - L| < ε
> >
> > In this age of unicode when we have xetex/luatex why do we use the first?
>
> Well over a decade ago, I went to a considerable amount of trouble
> (building on work of Renaud Marlet) to build a mode for (X)Emacs that
> let me see the latter (with simple compose-key inputs) instead of the
> former.

Thats a different direction.

In the days when the world started from one shore of the Atlantic
and ended at the other – the era of supremacy of ASCII – this made sense.
Just as it did for Knuth to encode all of Tex as ASCII.

Then the world grew larger with various latin-xx (iso-8859-xx)
And further fragmented with the windows codepages vs the above.
And finally people decided that the space inefficiency of a larger charset
is a minor headache compared to dozens of incompatible non-invertible
encodings.

Which saw the rise of unicode as a standard.
And among the many forward looking unicode-apps is xetex/luatex(?)

> I used it for a couple of years, and then gave up (though some of the
> work found its way into the x-symbol package, which I believe is still
> around).
>
> Why? It's just not worth the effort. If you speak LaTeX, it's hardly
> any easier to read the latter than the former; and when you really
> want to read it for content, you read the typeset output, whether on
> screen or on paper.

Well the start of this thread is a paper arguing the opposite!
Of course that argument may be flawed (and dishonest) but it contains
some truth viz that the encoding depth needed for tex functionality
is less in 2015 than in 1980.

[Luis Rivera]

On Monday, 5 January 2015 20:24:22 UTC-6, Rusi Mody wrote:
> On Tuesday, January 6, 2015 7:36:56 AM UTC+5:30, Luis Rivera wrote:
> > The alternative proposed somewhere else in this thread is XML. That is wrong too, imnsho: for XML is designed for machine and not human consumption. If you ask me, the least intrusive approach for an author is the use of something like WikiMarkup: a whole Encyclopedia is being edited by the second, as we speak, on this tiny markup. Leave the details of page layout or hyphenation to the publishers; whether they use HTML5, Plain TeX or InDesign shouldn't matter.
>
> +1 from me -- with the following caveat.
> Here is the wikipedia page on (ε,_δ)-definition_of_limit
> https://en.wikipedia.org/wiki/%28%CE%B5,_%CE%B4%29-definition_of_limit
>
> Editing it produces this excerpt
>
> (\forall \varepsilon > 0)(\exists \ \delta > 0) (\forall x \in D)(0 < |x - c | < \delta \ \Rightarrow \ |f(x) - L| < \varepsilon)
>
> Now compare it with
>
> ∀ ε > 0 (∃ δ > 0) (∀ x ∈ D) (0 < |x - c| < δ ⇒ |f(x) - L| < ε
>
> In this age of unicode when we have xetex/luatex why do we use the first?
>

Now try to typeset the following only with unicode.

:<math> \frac{a+b}{a} = \frac{a}{b} \ \stackrel{\text{def}}{=}\ \varphi,</math>

:<math>\varphi = \frac{1+\sqrt{5}}{2} = 1.6180339887\ldots.</math>

Nothing prevents you to write simple expressions directly in unicode, but it is far better to have a more general approach.

Cheers,

Luis.

[Michael Shell]

On Mon, 5 Jan 2015 18:24:18 -0800 (PST)
Rusi Mody <rusto...@gmail.com> wrote:

> (\forall \varepsilon > 0)(\exists \ \delta > 0) (\forall x \in D)(0 < |x - c | < \delta \ \Rightarrow \ |f(x) - L| < \varepsilon)
>
> Now compare it with
>
> ∀ ε > 0 (∃ δ > 0) (∀ x ∈ D) (0 < |x - c| < δ ⇒ |f(x) - L| < ε
>
> In this age of unicode when we have xetex/luatex why do we use the first?
>
> Hopefully most people would agree the latter is more readable than the
> former.

Readability is only one side of the coin. The other side is
*write/createability*, - the problem of identification and initial
entry. Given that I have only ten fingers and a hundred or so keys
in front me, how am I to invoke a specific symbol from the hundred
thousand or so that are available in Unicode?

In the above example, what if someone is not familiar with the
appearance of varepsilon, or if there are two or more complex symbols
that are easy to visually mistake for one another?

Are we to have to rely on cutting and pasting from preexisting works
(assuming they exist)? Or go through code pages trying to find the
correct key sequences or unmemorable Unicode numbers?

I can, however, easily remember descriptive names such as \forall.
And this is the whole underlying point of Western languages - that we
build all words out of sequences of a small and very managable set of
letters. We do not have to know the meaning of a word (or, likewise,
all the complexities of the visual appearance of a glyph) to be able
to easily reproduce (and remember) it.

Historically, the Western approach to language has shown itself to be
much more amenable to use with machines than the pictogram approach
used in the Far East.


Cheers,

Mike Shell

[Joost Kremers]

Michael Shell wrote:
> On Mon, 5 Jan 2015 18:24:18 -0800 (PST)
> Rusi Mody <rusto...@gmail.com> wrote:
>
>> (\forall \varepsilon > 0)(\exists \ \delta > 0) (\forall x \in D)(0 < |x - c | < \delta \ \Rightarrow \ |f(x) - L| < \varepsilon)
>>
>> Now compare it with
>>
>> ∀ ε > 0 (∃ δ > 0) (∀ x ∈ D) (0 < |x - c| < δ ⇒ |f(x) - L| < ε
>>
>> In this age of unicode when we have xetex/luatex why do we use the first?
>>
>> Hopefully most people would agree the latter is more readable than the
>> former.
>
>
> Readability is only one side of the coin. The other side is
> *write/createability*, - the problem of identification and initial
> entry. Given that I have only ten fingers and a hundred or so keys
> in front me, how am I to invoke a specific symbol from the hundred
> thousand or so that are available in Unicode?

[...]

> Are we to have to rely on cutting and pasting from preexisting works
> (assuming they exist)? Or go through code pages trying to find the
> correct key sequences or unmemorable Unicode numbers?
>
> I can, however, easily remember descriptive names such as \forall.

So use an editor that gives you the option to type `\forall' and then
inserts ∀ for you. :P

(Actually, for complex math formulae, I'm not sure using Unicode is the
best idea. Personally, I only use the occasional simplistic logic
formula, which can be handled just fine using Unicode, so I generally do.)

> Historically, the Western approach to language has shown itself to be
> much more amenable to use with machines than the pictogram approach
> used in the Far East.

That's actually not a property of the language but of the writing
system, which, although intimately connected to the language it records,
is a separate thing.



--
Joost Kremers joostk...@fastmail.fm
Selbst in die Unterwelt dringt durch Spalten Licht
EN:SiS(9)

[Rusi Mody]

Sure the inherent non-linearity of math cannot be handled by unicode – sums, integrals, limits, matrices…
Nor the beauty of typeset material.

The point is not latex OR unicode but unicode-tex or ASCII-tex.
So the frac would remain (and the math)

and so

:<math>\varphi = \frac{1+\sqrt{5}}{2} = 1.6180339887\ldots.</math>

should become something like
<math> ϕ = \frac{1+√{5}}{2} = 1.6180339887… .</math>

[Rusi Mody]

On Wednesday, January 7, 2015 2:40:05 AM UTC+5:30, Michael Shell wrote:

> I can, however, easily remember descriptive names such as \forall.
> And this is the whole underlying point of Western languages - that we
> build all words out of sequences of a small and very managable set of
> letters. We do not have to know the meaning of a word (or, likewise,
> all the complexities of the visual appearance of a glyph) to be able
> to easily reproduce (and remember) it.
>
> Historically, the Western approach to language has shown itself to be
> much more amenable to use with machines than the pictogram approach
> used in the Far East.

Sure.

But not sure how that relates to
(\forall \varepsilon > 0) etc

vs

∀ ε > 0 etc
Are you regarding ∀ ε as pictograms?
They are not found on keyboards aof course – eastern or western…
See below

> Readability is only one side of the coin. The other side is
> *write/createability*, - the problem of identification and initial
> entry. Given that I have only ten fingers and a hundred or so keys
> in front me, how am I to invoke a specific symbol from the hundred
> thousand or so that are available in Unicode?
>

ASCII is too small – so we need tables in WYSIWYG editors or ASCII encodings
like \varphi \epsilon etc in Latex

Unicode is too large – we need to use only what we need.
The suggestion is not to go from inputting exclusively-ASCII to
full unicode but to go from ASCII to ASCII + math.

Probably going from ½ % unicode-use to 1%.

Multilingual text is another question altogether

> In the above example, what if someone is not familiar with the
> appearance of varepsilon, or if there are two or more complex symbols
> that are easy to visually mistake for one another?
>
> Are we to have to rely on cutting and pasting from preexisting works
> (assuming they exist)? Or go through code pages trying to find the
> correct key sequences or unmemorable Unicode numbers?
>

Its probably good to distinguish these levels of input methods

1. Cut paste a character after running google
2. Select a character from a local input-method app like ibus
3. Use an editor input method eg emacs' tex input-method will convert
\forall into ∀ etc
4. Use the compose-key
5. Switch keyboard layouts in software
6. Use a special purpose hardware keyboard

As we go from 1 to 6 the expertise and efficienty increases but also
the expense of setup, hardware etc.

Around 3-4 is a quite good combination in efficiency and easiness

[Michael Shell]

On Tue, 6 Jan 2015 19:01:41 -0800 (PST)
Rusi Mody <rusto...@gmail.com> wrote:

> Are you regarding ∀ ε as pictograms?

Yes, although I now see "glyph" or "ideo/logogram" would have been a
better term to use. Under Unicode, they are essentially a specific
visual representation of a 32bit number (and vice versa). It would
solve everything if people could work directly with the raw numbers,
but this is not practical.

The question then becomes: what is the best way for a person
(not machine-to-machine which can use numbers directly) to tell
a machine which symbol we want? My position is that the best approach
is to use a sequence of ASCII characters as a "name" for the symbol
number (e.g., \varepsilon) due to the physical limitations of the
number of keys on a keyboard, and human memory.

A person who knows Greek will understandably want to work directly
with Greek glyphs, but I think most others would prefer to call the
symbols by an ASCII name. This may be true whenever symbols are
"foreign" to a given individual and not directly represented on
his/her keyboard.

Ironically enough, even with its limitations, I think ASCII is a
good standard on which to base names of symbols in an international
context. (e.g., a Chinese speaker dealing with Arabic glyphs)

> Its probably good to distinguish these levels of input methods
>
> 1. Cut paste a character after running google
> 2. Select a character from a local input-method app like ibus
> 3. Use an editor input method eg emacs' tex input-method will convert
> \forall into ∀ etc
> 4. Use the compose-key
> 5. Switch keyboard layouts in software
> 6. Use a special purpose hardware keyboard

I think that well summarizes the options for that route.
- vs. inputting \forall


Cheers,

Mike

[Torsten Bronger]

Hallöchen!

Rusi Mody writes:

> [...]

>
> The point is not latex OR unicode but unicode-tex or ASCII-tex.
> So the frac would remain (and the math)
>
> and so
> :<math>\varphi = \frac{1+\sqrt{5}}{2} = 1.6180339887\ldots.</math>
>
> should become something like
> <math> ϕ = \frac{1+√{5}}{2} = 1.6180339887… .</math>

See http://www.unicode.org/notes/tn28/UTN28-PlainTextMath.pdf. TeX
still has better typography and more features, but Word has long
supassed LaTeX in ease and readability of its math input syntax.

[Joseph Wright]

On 07/01/2015 08:02, Torsten Bronger wrote:
> Hallöchen!
>
> Rusi Mody writes:
>
>> [...]
>>
>> The point is not latex OR unicode but unicode-tex or ASCII-tex.
>> So the frac would remain (and the math)
>>
>> and so
>> :<math>\varphi = \frac{1+\sqrt{5}}{2} = 1.6180339887\ldots.</math>
>>
>> should become something like
>> <math> ϕ = \frac{1+√{5}}{2} = 1.6180339887… .</math>
>
> See http://www.unicode.org/notes/tn28/UTN28-PlainTextMath.pdf. TeX
> still has better typography and more features, but Word has long
> supassed LaTeX in ease and readability of its math input syntax.
>
> Tschö,
> Torsten.

Interesting paper, but at something of a tangent in some ways. There is
a reasonable amount there about using Unicode input and automatically
interpreting as maths, so for example the input

(a + b)/c

giving in (plain) TeX terms

$${a + b} \over c $

or in LaTeX

\[ \frac{a + b}{c} \]

That's a slightly different thing to the business of allowing √ directly
for \sqrt. Indeed, the paper says

> However unlike those formats, it doesn’t attempt to include all
> typographical embellishments.

One thing perhaps that is worth noting there is that the author
discusses input of the various symbols required, and suggests that at
least in some cases using something akin to TeX input is desirable

> This approach is noticeably faster than using menus ...

where an editor-based substitution ('AutoCorrect' for Word) is used to
give the real symbol in the file. This part can happily be done using a
Unicode TeX engine and suitable editor set up: it's the
'transformations' part that is more tricky (though almost certainly
doable either at the macro (XeTeX/LuaTeX) or Lua (LuaTeX) level).
--
Joseph Wright

[Joseph Wright]

Another thing that remains tricky with such an approach is dealing with
variables and operations, for example "sin(x)". The paper mentions this
but doesn't quite explain (to me) the outcome in terms of approach. They
mention you can set up to map the "x" here to the math alphabet block
('letter-like') and leave the "sin" as letters. That's a tricky one:
it's not necessarily going to be obvious to the user!

Overall, I think any sensible reading here does suggest that a 'mixed'
approach ('unicode-tex') is the best way to imagine handling maths.
--
Joseph Wright

[Joseph Wright]

On 07/01/2015 08:39, Joseph Wright wrote:

> On 07/01/2015 08:02, Torsten Bronger wrote:
>> Hallöchen!
>>
>> Rusi Mody writes:
>>
>>> [...]
>>>
>>> The point is not latex OR unicode but unicode-tex or ASCII-tex.
>>> So the frac would remain (and the math)
>>>
>>> and so
>>> :<math>\varphi = \frac{1+\sqrt{5}}{2} = 1.6180339887\ldots.</math>
>>>
>>> should become something like
>>> <math> ϕ = \frac{1+√{5}}{2} = 1.6180339887… .</math>
>>
>> See http://www.unicode.org/notes/tn28/UTN28-PlainTextMath.pdf. TeX
>> still has better typography and more features, but Word has long
>> supassed LaTeX in ease and readability of its math input syntax.
>>
>> Tschö,
>> Torsten.

[snip]

> This part can happily be done using a
> Unicode TeX engine and suitable editor set up: it's the
> 'transformations' part that is more tricky (though almost certainly
> doable either at the macro (XeTeX/LuaTeX) or Lua (LuaTeX) level).

What isn't said is that this 'inline' approach requires that the entire
expression is parsed before any output is produced. Taking the example

\varphi = \frac{1+\sqrt{5}}{2} = 1.6180339887\ldots

which presumably becomes

ϕ = (1 + √5)/2 = 1.6180339887…

getting the fraction right requires that once the "/" is spotted the
part before it is collected up (the opening bracket I don't think is
enough). There could of course be more of these cases, so you have to
look ahead. I guess that means internally any implementation of the
algorithm ends up needing internally something akin to the TeX version
to actually get the display right!
--
Joseph Wright

[Rusi Mody]

On Wednesday, January 7, 2015 1:36:06 PM UTC+5:30, Torsten Bronger wrote:
> Hallöchen!
>
> Rusi Mody writes:
>
> > [...]
> >
> > The point is not latex OR unicode but unicode-tex or ASCII-tex.
> > So the frac would remain (and the math)
> >
> > and so
> > :<math>\varphi = \frac{1+\sqrt{5}}{2} = 1.6180339887\ldots.</math>
> >
> > should become something like
> > <math> ϕ = \frac{1+√{5}}{2} = 1.6180339887… .</math>
>
> See http://www.unicode.org/notes/tn28/UTN28-PlainTextMath.pdf. TeX
> still has better typography and more features, but Word has long
> supassed LaTeX in ease and readability of its math input syntax.

Interesting!
And a module for converting this to standard (xe|lua)latex
does not yet exist? Should not be too hard to write...

[John Harper]

Torsten Bronger wrote:

> See http://www.unicode.org/notes/tn28/UTN28-PlainTextMath.pdf. TeX
> still has better typography and more features, but Word has long
> supassed LaTeX in ease and readability of its math input syntax.

Maybe, but when I go to talks by people whose slides used Word I am usually
horrified by the nasty appearance of their mathematics, e.g. different fonts
for the same symbol in different places and poor spacing, in particular. I
also recall going to a seminar in Cambridge UK where many Greek letters were
still Greek letters, but the wrong ones. The speaker said sadly "It was all
right in London." He had used Word.

--
John Harper

[Robert Heller]

MS-Word is *notorious* making things look *different* on different computers.
Mainly because a MS-Word file is not finally typeset -- it gets *freshly*
typeset when you load it into the local machine's incarnation of MS-Word and
is then subject to the local machine's setup (O/S version, installed fonts,
display settings, etc.). My guess is the the speaker brought his slides on the
thumb drive and the computer in Cambridge UK had some *subtle* differences
from the computer in London. With LaTeX you would generally transport the
*PDF* file -- a PDF file is effectively a stone tablet -- all of the
typesetting is done and if all of the fonts are embedded, there is no
possiblity of things being messed with by the local machine's setup. Note:
even if he transported the LaTeX source, TeX / LaTeX is much more standardized
as to how things are typeset. Less subject to software designer (or the
marketing department) whims.

[Joost Kremers]

John Harper wrote:
> also recall going to a seminar in Cambridge UK where many Greek letters were
> still Greek letters, but the wrong ones. The speaker said sadly "It was all
> right in London." He had used Word.

I've seen the same happening with LaTeX-based presentations. There used
to be a time when LaTeX didn't automatically embed fonts in pdfs and
this sort of thing wasn't uncommon. (I think nowadays most TeX
distributions are set up to embed fonts by default, though.)

It's actually possible to embed fonts in a Word document as well, but
I'm not sure if it's the default or not.

[Luis Rivera]

On Wednesday, 7 January 2015 15:41:24 UTC-6, Joost Kremers wrote:
> John Harper wrote:
> > also recall going to a seminar in Cambridge UK where many Greek letters were
> > still Greek letters, but the wrong ones. The speaker said sadly "It was all
> > right in London." He had used Word.
>
> I've seen the same happening with LaTeX-based presentations. There used
> to be a time when LaTeX didn't automatically embed fonts in pdfs and
> this sort of thing wasn't uncommon. (I think nowadays most TeX
> distributions are set up to embed fonts by default, though.)
>
> It's actually possible to embed fonts in a Word document as well, but
> I'm not sure if it's the default or not.
>

In these cases (when you have to do a presentation), you are the actual "publisher"; there you have full responsibility, and consequently, you need full control. In this case, the best output format is certainly PDF with embedded fonts (here PPT and ODP equally st!nk, just as DOCX and ODT in the word processing area do). Here, in terms of quality, *TeX* still rules; now, whether is it Plain, LaTeX or ConTeXt (or any other macro package) is still matter of debate.

What is at stake though is which one of these document formats (ppt, odp, tex, latex, context) is better suited for human production and long term storage, retrieval or recycling. I'd say, following a hint by Philip Taylor, that in these terms *TeX*, in its actual avatars, is as bad as any WordProcessing software. See

http://tug.org/interviews/taylor.html

Check especially the mistakes made by others pointed out by Taylor at the end of the interview. The markup language[s] is the first thing that needs to be fixed from the viewpoint of the end user; and I doubt *TeX* does nothing to fix it.

Cheers,

Luis.

[Peter Flynn]

On 01/01/2015 07:26 PM, Khaled Hosny wrote:
> I know a major scientific publisher that does not even accept LaTeX
> files from authors, yet the printed version is eventually typeset
> with LaTeX.

With good reason: most "LaTeX" submitted by authors is ghastly, and
requires many hours of editing to make it usable. Authors fail to
include the homebrew and weirdass packages they use, and they have a
particular liking for \defining macros for commonly-used constructs
which end up conflicting with control sequences defined elsewhere, or
which spew error messages.

In one book we typeset a few years ago, the authors had wrapped every
equation or group of consecutive equations in a \resizebox designed to
force the block to fit the margins. This meant that almost every
equation was a different type size; the authors said they didn't feel it
was important (or more likely had never noticed, because I now believe
that this wrapping was actually done under the hood by some WYSIWYG
interface in their editor). Every single one of these had to be undone,
and the equations (if too wide) rearranged to fit properly. Fortunately
we were being paid by the publisher to do the job, otherwise we would
have returned it with a polite message saying that the code was not usable.

A few hardy souls turn in clean, workable code that processes error-free
first time, but these are a minority.

Publishers have grown to mistrust LaTeX for this reason (wrongly: it's
the authors they should mistrust, not the software).

///Peter

[Peter Flynn]

On 12/28/2014 06:27 AM, Scott Pakin wrote:
> On 12/27/2014 11:46 AM, Torsten Bronger wrote:
>>> I'm not sure as far as LaTeX is concerned, but TeX remains since
>>> 1982 (more than 30 years) the leading typesetting system for
>>> science, and will for the next decades for sure.
>>
>> Is there any solid evidence for this?

No, I think it's nonsense. Most of the scientists I deal with, in my own
institution and elsewhere, have never heard of it: or they might
remember having seen it when they were doing their graduate work, but
they now use Word because that's what their institution provides.

Exceptions: obviously, mathematicians, physicists, engineers, and others
who have heavy-duty math work.

> [...] Hence, my guess
> is that one could not find solid evidence for TeX or LaTeX being the
> "leading typesetting system for science", but also that one probably
> could not find solid evidence that Word serves that role, either.

Numerically Word must by now outnumber LaTeX by several orders or magnitude.

> Personally, I'm curious what typesetting systems the major scientific
> publishers use internally. They all claim to use proprietary systems,
> but I would guess that they're really using (La)TeX or Word plus a
> bunch of in-house macros -- whether the copy editors realize that or
> not.

No, most of them use InDesign, Quark XPress (still), or (for those using
more XML) FrameMaker or Arbortext Publisher (or whatever PTC is calling
it this week). Some might still use 3B2, a venerable typesetter with a
curious backslash-and-curly-brace syntax (wonder where they got that
from :-) that was bought up by Adobe some years ago but is still used.
There are still journals using older software like PageMaker (my
brother's house still uses Ventura Publisher) simply because of their
stability. A very small number use LaTeX to final production: Kluwer,
for example, and parts of Springer and Elsevier, and smaller journals
handling material with heavy math content.

As Joseph pointed out, TeX once formed the typesetting engine inside
Arbortext Publisher (Arbortext actually sold their own TeX system as a
product once); it also formed the core code around which other systems
were written, such as 3B2 and some parts of Quark XPress.

///Peter

[Joost Kremers]

Luis Rivera wrote:

> What is at stake though is which one of these document formats (ppt,
> odp, tex, latex, context) is better suited for human production and
> long term storage, retrieval or recycling. I'd say, following a hint
> by Philip Taylor, that in these terms *TeX*, in its actual avatars, is
> as bad as any WordProcessing software. See
>
> http://tug.org/interviews/taylor.html
>
> Check especially the mistakes made by others pointed out by Taylor at
> the end of the interview. The markup language[s] is the first thing
> that needs to be fixed from the viewpoint of the end user; and I doubt
> *TeX* does nothing to fix it.

I can't say I find the argument very convincing. The only mistake Taylor
seems to point out is when users define a macro for a frequently-used
word and then forget to put a backslash after it to prevent
space-gobbling. I would argue that abbreviating frequently-used words
should be handled by the editor anyway (e.g., Emacs' abbrev-mode or
TextExpander on Mac/iOS; Word and LibreOffice can do this as well).

I also see no advantage in an HTML-like syntax. IMHO it's far too
verbose, requiring each tag twice. I think the advantage that it has
(having to close a tag explicitly) can be achieved in LaTeX fairly
easily by requiring that each macro have a pair of braces. (Much like
many programming languages do, where «myfunc» is simply an identifier
referring to a function object, while «myfunc()» is the way to actually
call the function.)

This would mean that switches such as \bf are not allowed anymore
(rather than being "obsolete"), and that *if* you want to define a macro
to abbreviate frequently-used words, you need to use it as «\LC{}»
rather than «\LC». That would break backward compatability, but that's
been done before.

[Joseph Wright]

On 07/01/2015 23:15, Peter Flynn wrote:
>> Personally, I'm curious what typesetting systems the major scientific
>> publishers use internally. They all claim to use proprietary systems,
>> but I would guess that they're really using (La)TeX or Word plus a
>> bunch of in-house macros -- whether the copy editors realize that or
>> not.
>
> No, most of them use InDesign, Quark XPress (still), or (for those using
> more XML) FrameMaker or Arbortext Publisher (or whatever PTC is calling
> it this week). Some might still use 3B2, a venerable typesetter with a
> curious backslash-and-curly-brace syntax (wonder where they got that
> from :-) that was bought up by Adobe some years ago but is still used.
> There are still journals using older software like PageMaker (my
> brother's house still uses Ventura Publisher) simply because of their
> stability. A very small number use LaTeX to final production: Kluwer,
> for example, and parts of Springer and Elsevier, and smaller journals
> handling material with heavy math content.

Things might be slightly different for books from journal articles: I've
been involved recently in a book chapter for Wiley, and the proofs there
clearly where generated using TeX. On the other hand, their chemistry
journals are set using ArborText. So it's a mixed situation even within
one publishing house. That said, I'd agree that most of the journal
articles I see are not typeset using TeX at the 'business end'.
--
Joseph Wright

[Joseph Wright]

With my 'LaTeX3 hat' on this sort of thing is something to think about.
As you say, HTML is more verbose than (La)TeX over all, and if one goes
toward XML that is even more the case. There's an argument for going
slightly in that direction (see ConTeXt MkIV's preferred treatment of
sections and list items), but likely not for everything.

Suggestion about requiring an argument in all cases is interesting
although might be more problematic for cases where the space skipping is
useful

{\textbf Some text} =>
{\textbf{}Some text} % Can't have a space before "S"

(I'm not really sure what he means about backslashes and braces: they
are just some selection of special chars in the input. HTML has < and &
(at least) as special, so it comes down as much to convention as
anything else.)

What I think we can agree is needed is more clarity on the division
between logical and appearance markup, and more flexible ways of
altering the behaviours. A note of caution of course is that TeX works
for typesetting as it does allow some visual adjustments. The issue with
a 'purely logical markup' approach is that there are edge cases that
need intervention manual. Thus even in an XML world there needs to be
the odd equivalent to \vskip, \-, etc.
--
Joseph Wright

[Eric Pozharski]

with <slrnmara00.jq7....@j.kremers4.news.arnhem.chello.nl>
Joost Kremers wrote:

*SKIP*

> It's actually possible to embed fonts in a Word document as well, but
> I'm not sure if it's the default or not.

It's off. As usual with MSW's features nobody uses it, so, probably,
Redmond doesn't know the feature doesn't work. I've done this
font-embedding (OMG) twenty years ago. It was a catstrophe.

p.s. However, I have an option to try it again (without such
devastating blow to my reputation).

--
Torvalds' goal for Linux is very simple: World Domination
Stallman's goal for GNU is even simpler: Freedom

[Axel Berger]

Luis Rivera wrote on Wed, 15-01-07 23:14:

>that in these terms *TeX*, in its actual avatars, is as bad as any
>WordProcessing software.

That may be so on first opening (compiling) an old file. When it comes
to repairs I expect an ASCII source to be head, shoulders, and hips
above undocumented binary.

[Marc van Dongen]

On Monday, 5 January 2015 21:12:06 UTC, jfh wrote:
> Marc van Dongen wrote:
>
> > On Monday, 5 January 2015 13:19:03 UTC, Robert Heller wrote:
> >
> >> Right. The paper pretty obviously played all of the *same* sorts of
> >> 'games'
> >> early Linux vs. MS-Windows FUD studies played. The study was
> >> *intentionally*
> >> rigged to make MS-Word 'look better', using a non-real world case. It
> >> was an Apples vs. Oranges comparison where the goal was to test for the
> >> more orange fruit.
> >
> > I forgot to mention one thing. The displayed equations in the tests in the
> > paper don't have any end-of-sentence punctuation symbols, which they
> > _should_ have. It clearly demonstrates the paper is about being good at
> > copying/reproducing/typesetting a given piece of text, not about doing it
> > right!
>
> Some good authors (or editors?) disagree with you. Two Cambridge University
> Press books on fluid mechanics have: punctuation after almost all displayed
> equations (Drazin & Reid), and after almost none (Pozrikidis).

If that's their style, I am not sure why you call them good:-). A displayed equation is part of the sentence and if it's at the end of the sentence, it requires an end-of-sentence punctuation symbol.

I thought _Mathematical Writing_ said something about it but I coudn't find the reference. (Perhaps I'm wrong.) Anyway, when I googled for it, I came across many lecutre notes/papers that seemed to agree with me.

Regards,


Marc van dongen

[Julian Bradfield]

On 2015-01-08, Marc van Dongen <don...@cs.ucc.ie> wrote:
[ snip ]

> If that's their style, I am not sure why you call them good:-). A displayed equation is part of the sentence and if it's at the end of the sentence, it requires an end-of-sentence punctuation symbol.

On the other hand, a displayed equation is itself a cohesive
mathematical entity, and doesn't deserve to be polluted by a
non-mathematical punctuation sign (especially if ! or ?).

Clearly, the logical thing is to show the equation, followed by
punctuation in the place where the sentence continues. So, for
example, my favourite equation is

$$\mathrm{e}^{i\pi} + 1 = 0$$

. It's just unfortunate that that looks a bit odd:-)

[Marc van Dongen]

On Thursday, 8 January 2015 15:05:09 UTC, Julian Bradfield wrote:

That's all fine, but you always want to know whether such entity is the end of a statement/sentence. As an extreme case, assume your display is at the bottom of a recto page. Should that happen, you simply cannot tell and have to look at the next page first, which is extremely awkward.

In a similar vein, consider the following: ``Since $a = b$ and $b = c$, we have $a = c$.'' In displays-are-sentences-in-their-own right we could write ``\[ a = b \]\[b = c\] Therefore, \[a = c\]'' There's no real difference between a display and an inline formula, so let's turn the displays into inline math. We now get ``$a = b$ $b = c$. Therefore, $a = c$'' Somebody may _parse_ this as ``$a = b b = c$. Therefore, $a = c$'' Clearly, this should not be allowed.

Regards,


Marc van Dongen

[Luis Rivera]

On Thursday, 8 January 2015 01:40:16 UTC-6, Joost Kremers wrote:
> Luis Rivera wrote:
>

> > Check especially the mistakes made by others pointed out by Taylor at
> > the end of the interview. The markup language[s] is the first thing
> > that needs to be fixed from the viewpoint of the end user; and I doubt
> > *TeX* does nothing to fix it.
>
> I can't say I find the argument very convincing. The only mistake Taylor
> seems to point out is when users define a macro for a frequently-used
> word and then forget to put a backslash after it to prevent
> space-gobbling.

The main mistake pointed out by Taylor is that TeX's default syntax (and perhaps the whole design, I might add) is prone to allow the user to mix up logical markup with page layout with macro programming. He proposes markup akin to HTML, and I don't quite follow him on that, as stated elsewhere; what I agree with is that authors should be taken away the ability to control minor points of display (hyphenation and page breaks, for instance) to allow them to concentrate in the contents.

Think of the way many webpages work nowadays: the main content may be written in HTML-4, some aspects of the layout are written in CSS, and some programming within the page is written in JavaScript. There you have the three aspects clearly distinguished by language.

> I also see no advantage in an HTML-like syntax. IMHO it's far too
> verbose, requiring each tag twice. I think the advantage that it has
> (having to close a tag explicitly) can be achieved in LaTeX fairly
> easily by requiring that each macro have a pair of braces. (Much like
> many programming languages do, where «myfunc» is simply an identifier
> referring to a function object, while «myfunc()» is the way to actually
> call the function.)
>

As I said above, I do not advocate [x|ht]ml in particular, though there is experience that something like that may be useful for beginners. StarTeX has some HTML flavor in it, and it succeeds in standardizing the users' markup and output, leaving out the details of the document formatting and allowing beginners to concentrate in their topics.

Consider, by contrast, Texinfo: Texinfo has a very sparse syntax (only @{} are significant), and it doesn't have the redundance of [X|HT]ML; you can write user macros, you can produce HTML, DVI/PDF or XML on demand, and you can tell the system what to do for a particular target format.

Both have their shortcomings, the main one being that the DVI/PDF output is pretty much hardcoded; but they also show that using *TeX* as a typesetting engine doesn't prevent the specification of a simpler markup language which avoids the mixup of functionality allowed by pure *TeX* macros.

I wish LaTeX3 could specify a markup like Texinfo's: minimalistic yet easy to convert to other formas (OOXML, ODF, perhaps), so that there would be three levels of development: user's contents, macro programmers, and core maintainers. Only the last would have to deal with *TeX* macros, and users would clearly realize when they are meddling with stuff they would not really worry about if they are not the final typesetters.

Cheers,

Luis.

[Scott Pakin]

On 01/08/2015 09:11 AM, Marc van Dongen wrote:
> That's all fine, but you always want to know whether such entity is the end of a statement/sentence.

I think you missed the period that Julian put at the beginning of the
*next* line:

> On Thursday, 8 January 2015 15:05:09 UTC, Julian Bradfield wrote:
>> . It's just unfortunate that that looks a bit odd:-)

^
|
+-- this one

I agree that that looks odd and have never encountered a style guide
that promoted (or even accepted) such typesetting, even though it does
clarify whether the equation ends a sentence.

In my own work, I include the punctuation in the displayed equation, but
I stick in a \qquad or other big space to clearly separate the
mathematics from the punctuation. It's still ugly, but at least the
meaning is unambiguous.

Anecdote: I remember a friend of mine in college trying to make sense of
an equation in his abstract-algebra textbook -- something like "a = b^2".
He couldn't figure out why the result was b^2 and not b. After spending
hours pondering and going through reams of scratch paper, he finally
realized that the superscript meant "footnote 2", not "b squared".
Groan!

-- Scott

[Luis Rivera]

Point conceded about undocumented binaries. What about _documented_ ones (OOXML, ODF, etc.)

And then, you restrict your choice to ASCII files; I'd rather include many more plain text formats (that would include UTF8 encoded files).

Cheers,

Luis.

[Luis Rivera]

On Wednesday, 31 December 2014 04:37:36 UTC-6, Dr Eberhard Lisse wrote:
> Some arguments about useability, such as spelling errors and such make
> sense.
>
> But LyX is an answer.
>

LyX has the drawback of relying on LaTeX, and thus it has all the shortcomings of the latter, plus some of its own. For instance, it allows the user to insert directly TeX commands (appropriately called Evil Red Text), and thus doesn't shield the end user from direct exposure to problematic TeX control sequences.

Cheers,

Luis.

[Karl]

That´s ridiculous. Nobody forces end users to _put_ TeX commands in
there. If you don´t like it, leave it and stick to the functionality LyX
provides.

[Torsten Bronger]

Hallöchen!

Karl writes:

> Am 09.01.2015 um 00:08 schrieb Luis Rivera:
>

>> [...]

>>
>> LyX has the drawback of relying on LaTeX, and thus it has all the
>> shortcomings of the latter, plus some of its own. For instance,
>> it allows the user to insert directly TeX commands (appropriately
>> called Evil Red Text), and thus doesn't shield the end user from
>> direct exposure to problematic TeX control sequences.
>
> That´s ridiculous. Nobody forces end users to _put_ TeX commands
> in there. If you don´t like it, leave it and stick to the
> functionality LyX provides.

Luis' assertion is not backed by evidence, but it certainly is not
outright ridiculous. Logical markup as well as LyX'es "WYSIWYM"
(... mean) need some patience and discipline. The LaTeX documents
of my collegues, however, contain a lot of finetuning. The desire
for full control is very common here.

[Joost Kremers]

Luis Rivera wrote:
> [...] what I agree with is that authors should be taken

> away the ability to control minor points of display (hyphenation and
> page breaks, for instance) to allow them to concentrate in the
> contents.

Yes, you're right of course. In principle, I agree as well, but I
believe there is one complication (see below).

> As I said above, I do not advocate [x|ht]ml in particular, though
> there is experience that something like that may be useful for
> beginners. StarTeX has some HTML flavor in it, and it succeeds in
> standardizing the users' markup and output, leaving out the details of
> the document formatting and allowing beginners to concentrate in their
> topics.

Yet no-one seems to be using StarTeX. (That's to say, I had never heard
of it before, and a Google search for it yields a lot of pages referring
to things non-TeX-related. The only TeX-related site in the search
results is the CTAN page for StarTeX. And that's in spite of the fact
that Google knows very well that I do the occasional TeX-related
search...)

Part of that is probably momentum, but part of it is something else, I
think: people *want* to be able to adjust the layout from time to time,
and a platform that doesn't give them that ability won't attract many
users.

> Consider, by contrast, Texinfo: Texinfo has a very sparse syntax (only
> @{} are significant), and it doesn't have the redundance of [X|HT]ML;
> you can write user macros, you can produce HTML, DVI/PDF or XML on
> demand, and you can tell the system what to do for a particular target
> format.

Texinfo does have @noindent, which isn't content-related...

> I wish LaTeX3 could specify a markup like Texinfo's: minimalistic yet
> easy to convert to other formas (OOXML, ODF, perhaps), so that there
> would be three levels of development: user's contents, macro
> programmers, and core maintainers. Only the last would have to deal
> with *TeX* macros, and users would clearly realize when they are
> meddling with stuff they would not really worry about if they are not
> the final typesetters.

Yes, that sounds good. But equally important, I think, is educating
users about the fact that they shouldn't even *want* to deal with
formatting. (I've seen LaTeX courses at universities taught by people
who weren't even aware of this...)

[Justin Thyme]

GL wrote:
> Le 27/12/2014 02:38, tsd a écrit :
>> "Given these numbers it remains an open question to determine the
>> amount of taxpayer money that is spent worldwide for researchers to
>> use LaTeX over a more efficient document preparation system, which
>> would free up their time to advance their respective field. Some
>> publishers may save a significant amount of money by requesting or
>> allowing LaTeX submissions because a well-formed LaTeX document
>> complying with a well-designed class file (template) is much easier to
>> bring into their publication workflow. However, this is at the expense
>> of the researchers' labor time and effort. We therefore suggest that
>> leading scientific journals should consider accepting submissions in
>> LaTeX only if this is justified by the level of mathematics presented
>> in the paper. In all other cases, we think that scholarly journals
>> should request authors to submit their documents in Word or PDF
>> format. We believe that this would be a good policy for two reasons.
>> First, we think that the appearance of the text
> is secondary to the scientific merit of an article and its impact to the
> field. And, second, preventing researchers from producing documents in
> LaTeX would save time and money to maximize the benefit of research and
> development for both the research team and the public."
>>
>> http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0115069
>>
>>
> I agree with M Ketterlin : this is a fucking hoax ! ;-)

>
> I'm not sure as far as LaTeX is concerned, but TeX remains since
> 1982 (more than 30 years) the leading typesetting system for science,
> and will for the next decades for sure.

Presumably LaTeX can be both awful and the best there is.

--
Sorrow in all lands, and grievous omens.
Great anger in the dragon of the hills,
And silent now the earth's green oracles
That will not speak again of innocence.
David Sutton -- Geomancies

[Axel Berger]

Luis Rivera wrote on Thu, 15-01-08 20:18:

> Texinfo has a very sparse syntax

> you can write user macros, you
> can produce HTML, DVI/PDF or XML on demand,

I don't quite see the problem. The LaTeX syntax between \begin and
\end{document} can and should be very sparse and simple. It should then
be possible to translate into all kinds of other formats, but you will
probably need to build and set up something whenever the class and the
document head changes.
TeX is an extremely powerful typesetting engine allowing everything
while Epub and HTML need to be primitive to be adaptable to the viewer.
So from first principles it will never be possible to translate LaTeX
to them completely. Most elements of LaTeX are meaningless in Epub,
take bottom placement of figures, so you will have to make do with a
subset.

[Torsten Bronger]

Hallöchen!

Axel Berger writes:

> Luis Rivera wrote on Thu, 15-01-08 20:18:
>
>> Texinfo has a very sparse syntax you can write user macros, you
>> can produce HTML, DVI/PDF or XML on demand,
>
> I don't quite see the problem. The LaTeX syntax between \begin
> and \end{document} can and should be very sparse and simple.

The problem is that many authors wouldn't have the discipline to
keep it sparse and simple. You may blame the authors -- but people
are as they are. However, programs can be changed.

15 years ago, I considered myself an author and an (amateur)
typesetter when it came to writing. But today, I only want to be an
author, so I do many things in org-mode and reStructuredText now. I
really like that lightweight markup *forces* me to keep things
simple. I even happily accept sub-optimal formatting here and
there.

[Luis Rivera]

I didn't mean to say that users are forced to insert TeX css in their documents; my point is that nothing prevents them to do so; so we are back to the point where authors are tempted to become book designers.

LyX alone (even without ERT) is still prone to that, given all the room for customizations it already has; but what really bugged me about the latest LyXes was that, in order to support all the "built-in" customizations, it _required_ me to install a lot of LaTeX packages I seldom if ever use. The first LyXes (which I used) were OK, but the last ones slightly slide on the side of bloatware.

Very opinionated,

Luis.

[Peter Flynn]

On 01/08/2015 02:30 PM, Marc van Dongen wrote:
[...]

> If that's their style, I am not sure why you call them good:-). A
> displayed equation is part of the sentence and if it's at the end of the
> sentence, it requires an end-of-sentence punctuation symbol.

I usually make one exception when setting text (not math) for print: if
a sentence ends with a URI, I do not add a period at the end, even with
a \thinspace, because it will probably mislead the user (and if the
author has provided one, I remove it). So far all the publishers
involved have agreed.

It's different for an online work (HTML, PDF) because the URI will be a
link; but from print, a user would have to type it. Stylistically, it's
probably not good to end a sentence with a URI, but that's an editorial
problem, not a typesetting one.

I can understand the perceived need to preserve the integrity of "a
sentence" when it ends with an expression, but aesthetically (not
mathematically) I think it looks very odd when a large displayed
equation has an orphan period after it. Having waded through deciphering
the equation and pondered on its content, it feels strange to be
accosted by a period and reminded (unnecessarily) that you were actually
reading a sentence :-) But I'm not a mathematician, and I remain guided
by mathematicians as to how they want their material typeset.

///Peter

[Peter Flynn]

On 01/08/2015 08:31 PM, Scott Pakin wrote:
> In my own work, I include the punctuation in the displayed equation, but
> I stick in a \qquad or other big space to clearly separate the
> mathematics from the punctuation. It's still ugly, but at least the
> meaning is unambiguous.

Ultimately, especially in a technical document, accuracy must always
trump aesthetics.

> Anecdote: I remember a friend of mine in college trying to make sense of
> an equation in his abstract-algebra textbook -- something like "a = b^2".
> He couldn't figure out why the result was b^2 and not b. After spending
> hours pondering and going through reams of scratch paper, he finally
> realized that the superscript meant "footnote 2", not "b squared".

A mathematician who would footnote an equation that way probably needs
retraining.

[This is, incidentally, the kind of error I was referring to when I
wrote about the poor quality of some authors' LaTeX.]

///Peter