On Fri, Nov 11, 2011 at 10:39 PM, A. Mani <a.mani....@gmail.com> wrote:
> (In French)
>
> http://fgallaire.flext.net/comparaison-langage-balisage-markup-lightweight-leger-txt2tags-pandoc-docutils-asciidoc-deplate-stx2any-aft-markdown-textile/
(My view: Automated mark up is never good enough.... and google
translate is imperfect in handling languages with solid structure (esp
if the target is English?). )
Google translate (tables figures missing):
Comparison of markup languages (markup) lightweight (lightweight):
Txt2tags, Pandoc, Docutils, AsciiDoc, Deplate, Stx2any, AFT, Markdown
and Textile
The office is the main use of computers since its inception. Yet the
majority of tools used in this field, software WYSIWYG word processor
such as OpenOffice and LibreOffice, leaving the majority of IT
professionals and ergonomists very doubtful, if not totally desperate.
These programs have in fact a very large number of defects: they focus
on form and not substance, their final result is often not what is
displayed, they are incompatible with each other, they are huge plants
Gas unusable on older, they only work in graphics mode, etc.. The only
rational, efficient and interoperable to work on a computer is to use
simple text files, all documents are therefore editable in any text
editor.
It was therefore necessary to think of a way to give these
instructions formatting in the text file itself, and thus appeared
markup languages (markup), the best known are HTML (invented in 1991
by Tim Berners-Lee) and LaTeX (established in 1985, and based on TeX,
invented by the great Donald Knuth in 1977), which was the first major
figure Roff, a Unix program history developed from 1961, which GNU,
Groff, is installed by default on all Linux distributions, since we
still use the man pages for software.
To better view, take as example the creation of a new section of a
document in man:
. SH New section in man
HTML:
<h1> New section in HTML </ h1>
and LaTeX:
\ Section {new section} LaTeX
These languages represent a significant improvement, but all have
one big problem: they are annoying! There are no longer content just
as easily in the middle of all these additional tags, not to mention
the fact that the complex syntax pave the way for many compilation
errors.
In 1995 we found the solution to this problem, with the creation of
the first language Wiki, whose main purpose was to allow easy editing
of web pages for everyone, and which the current user most famous is
the free encyclopedia Wikipedia. If there are almost as many different
syntaxes that Wiki software, they all have the characteristic of using
text characters simple and intuitive to give indications of text
formatting.
Always the same example, a new section in MediaWiki:
Section = = New Wiki
and one in Setext:
New section Setext
=======================
But why limit these lightweight markup languages in one generation
of HTML? Why not use the same syntax for different targets (called
backends, targets or as software writers), so as to obtain both a web
page in HTML, LaTeX a document for printing, or a page of man for
software? It is the software that is the aim that interest me, they
are for me the future of desktop computing, and I was led to compare
them to choose which one get involved as a developer.
Here is a comparison of the best existing free software, with
additional information as the existence of a target plain text,
because I wished to rely on the code to program some of my ideas ASCII
art.
The complete software
Name Popularity of Programming Languages Project License i18n Target
plain text
Txt2tags average Python doc + + Yes Yes RedNotebook GNU GPLv2
Python docutils strong Sphinx + No No Public Domain
AsciiDoc Python strong Dump No w3m or Lynx GNU GPLv2 or later
No Yes Ruby Deplate low GNU GPLv2 or later
Haskell average pandoc No Yes GNU GPLv2 or later
Sed and m4 low Stx2any Not Dump w3m personalized copyleft license
AFT No No small Perl Clarified Artistic License
Except Docutils and AFT (Almost Free Text), all software seemed at
first to propose a target text. But in fact it's a bit of a trompe
l'oeil, as two of them, and AsciiDoc Stx2any, simply dump a text-only
web browser (w3m or Lynx) income generated by the target HTML. There
was no in these cases the basic code that I could hope to improve.
I put in the comparative stx2any AFT and encoded respectively in Perl
and m4, so as to provide a wide range of programming languages, but
these programs are used both less and have fewer features than others.
By studying the different software available, I realized that,
fortunately for me, many were coded in Python (3 of 7 in total, but
above 3 on the 5 really interesting). Moreover, it is not the first
time that I see, trying to select free software, as coded in python
are both more numerous and better.
The five remaining programs are really excellent, and are all good
choices. Three of them, Docutils, Deplate Pandoc and have an advanced
design, with a finite state machine for which to write new readers and
writers so clean. However, despite their great qualities, is a project
too Deplate confidential (so it is not present among the
incomprehensibly yet so many Debian packages), and I felt not up to
get involved in a project like Pandoc, written entirely in Haskell,
which is a complex programming language that I would love to use, but
where my skill is still too limited.
It was necessary that I deepened my thinking about software in Python:
Choice in Python
Top Targets Target Name plain text
Txt2tags HTML, Latex and Wiki syntax Yes
Docutils HTML and Latex No
AsciiDoc and DocBook HTML Dump
No HTML Markdown
No HTML Textile
I added in this comparison Markdown and Textile, since they each have
an implementation in Python, but generate only the HTML, they do not
really interest me. AsciiDoc Txt2tags and have somewhat the same
architecture, with a big main file doing all the work that can be
configured respectively with a. Conf and two Python dictionaries (one
for the tags and the other for the Rules) to create new targets.
AsciiDoc Txt2tags and are therefore easier to learn and to change
quickly as Docutils, which is a very beautiful and well-architected
state machine object, but also more difficult to grasp.
Also, as I disapproved entirely the licensing policy of Docutils
public domain, it remained for me to make my choice Txt2tags and
AsciiDoc. It is mainly very DocBook orientation (one size does not
have a personal interest) of AsciiDoc and other details such as
location in many languages and Txt2tags its simplest, to me finally
made Txt2tags choose.
This choice is confirmed by further studies of different syntaxes. So
while the syntax is the Docutils has only:
* Italic * ** bold ** and
Txt2tags is much richer:
/ / Italic / / ** bold ** and __souligné__ - blocked -
Video encoding is much better, and understanding with instant syntax
Txt2tags since leaning slashes give the impression of italics, the
stars overhead mimic fat, underscores give the impression underscore,
and less appear as a bar. In addition, the widespread use of markup
characters in doubles, can lift at low cost a maximum of syntactic
ambiguities.
For example a sentence as simple as that contained in the sample file Txt2tags:
We use double *, /, - and _ to Represent ** bold **, / / italic
/ /, - strike - and __underline__.
and no problem:
is in fact too remains ambiguous, since:
We use double *, /, - and _ to Represent ** bold ** * italic *
underline and strike.
will:
In general, the syntax is often too heavy. For example if you insert
an image as the link:
.. picture: picture.png
: Target: http://fgallaire.flext.net
that can be compared with the much simpler Txt2tags:
[[Picture.png] http://fgallaire.flext.net]
We note here that the course provides syntax is certainly plenty of
other options that simple: target: and it is more powerful than
Txt2tags. However, my opinion is that the penalty of heaviness is
still there, although it is only useful in maybe 1% of cases.
Here is a comparison of software available syntax:
Sort by syntax
Use Markdown Txt2tags AsciiDoc reStructuredText
Office Txt2tags AsciiDoc Docutils
Pandoc Pandoc
Web server Txt2tags (Python) AsciiDoc (Python) Docutils (Python) Perl
Python
Ruby
PHP
Web Client JavaScript
Their implementation in Python allows Txt2tags, REST (for Docutils)
and AsciiDoc to be used as both office software platform (Linux, Mac
OS X, Windows and * BSD) and the web server side. Opposite, Markdown
is represented by an armada of implementations in all languages used
on the web server side, and also in client-side JavaScript for Ajax
preview without effective, but only Pandoc, which is not so easy to
compile on all platforms, offers something other than rendering HTML.
I will of course continue to work on the software Txt2tags, but an
implementation of the JavaScript syntax Txt2tags for a live report on
the net, as well as readers Pandoc because I really want to program in
Haskell, and Docutils for Then take advantage of the sublime Sphinx,
are projects that motivate me more.
Finally, I am always a little nostalgic at this screenshot because
it's seeing him, with the top left with the file tags, and the bottom
right one with the plain text result, I became aware Txt2tags that was
what I expected and more than it was in Python, it would probably be
the software that I would help!
Tags: AFT, Almost Free Text, AsciiDoc, lightweight markup, Office,
Deplate, Docutils, Haskell, lightweight markup language, markup
language, Free, Markdown, Markup, Pandoc, Python, REST,
ReStructuredText, Stx2any, Textile, Txt2tags, WYSIWYG
This entry WAS posted on Sunday, August 14th, 2011 at 3:51 and is
filed under Uncategorized. You Can Follow Any responses to this
entry-through the RSS 2.0 feed. You Can leave a response, or trackback
from your-own site.
10 Responses to "Comparison of markup languages (markup) lightweight
(lightweight): Txt2tags, Pandoc, Docutils, AsciiDoc, Deplate, Stx2any,
AFT, Markdown and Textile"
Shower Sébastien says:
15/08/2011 at 1:53
Hello Florent,
thank you for this post interesting, but I'm embarrassed about one
thing: you seem confused and markup language implementation. rst
example exists outside docutils.
I'm in the same thought in my company to replace LibreOffice for
the creation of our documents. And even if I find a tool more fun than
the Sphinx, the popularity and perinnité of the latter is strongly
sway the choice in favor.
Reply
fgallaire says:
15/08/2011 at 2:04
Yes, that separate the two that I made the final table, one
can see that a reader has Pandoc reST (even if it is less
comprehensive than Docutils).
Reply
Shower Sébastien says:
15/08/2011 at 2:58
rst2pdf is an interesting tool to make the doc:
http://code.google.com/p/rst2pdf/
Reply
fgallaire says:
15/08/2011 at 2:01 p.m.
rst2pdf is based on docutils he uses the 'reader' rst and the
state machine is a 'writer' pdf directly (without using LaTeX), thanks
to the use of the library ReportLab.
Reply
PetitPierre94400 says:
15/08/2011 at 3:54
Interesting comparison but for the advanced technical writing
(which must meet certain industry standards, for example), these
solutions do not offer enough opportunities.
@ Sebastian: For replacement of our "office", we use products from
Altova. I am fully aware that this is not free software but it is the
only fully meet our requirements and above all our needs and
expectations. The possibilities are such that we no longer have to go
into a DIY costly in time and money.
Reply
Elessar says:
16/08/2011 at 10:25
I want to say that worry is missing, but given the wealth of
alternatives, it would be difficult to list them all. I would like to
mention also Creole, which is a common wiki syntax. Nothing to do with
the horrible syntax MediaWiki course.
Reply
fgallaire says:
16/08/2011 at 7:49 p.m.
Cuckoo,
Creole is an interesting approach attempts to unify the wiki
syntax, I put in a link to the article thinking that they have done is
well documented and argued.
In fact they came to same conclusion as Txt2tags on a lot of
points. To quote my example of bold and italics:
http://www.wikicreole.org/wiki/BoldAndItalicsReasoning
"A star (*) IS The Most Used symbol to bold text online. A
slash (/) looks like slanted italics, so intuitive and It Is Easier To
remember Malthus. Double symbols are Generally Used in Creole to Avoid
accidentally parsing text not parsed Meant to Be. "
By cons there is no implementation of Creole that produces
something other than HTML ... So I do not think I forgot to software
that matches my comparison.
Reply
eric says:
21/08/2011 at 9:31 p.m.
Hi Florent,
With stiff competition in this niche there, it's still a chance
for txt2tags have you recovered as a developer!
Aurelio damage has not responded in relation to your request to
release a new version of txt2tags (2.7), as seen in the latest
development version that would be really a good time. No doubt it is
taken with many other projects, and the workload of doing that (and it
did for the 2.6) scares him. I am more in favor of "Release early,
release Often," not only because it can provide users with an updated
version of the latest features, but also because it allows to give an
excuse to "do the buzz" around, involving a dynamic event around the
release.
The mode of distribution txt2tags, a single file python + the doc,
that it is not very complicated to return to the previous version if
the user wishes (but a priori there is never really had a problem
compatibility with older files written in txt2tags).
An implementation of javascript txt2tags would be really great,
and also a php version, modeled on
http://michelf.com/projects/php-markdown/
Of course, not (necessarily) need to allow all that allows
txt2tags (macro, export in 30 formats), support for the syntax would
be great enough, and would bring txt2tags where already markdown
(wordpress plugins etc).
In short, as you indicated, the syntax txt2tags particular logic
seems to me one of his strengths, and if Creole is the right way, it
does not seem to move much on his side, so I find it txt2tags be worn
as far forward as possible, to be considered by more and more
projects, either syntax in text editors, or "wiki syntax" default
sites or discussion wiki online.
In this regard, I particularly appreciate the interface
RedNotebook that without wysiwyg, previews rendering (/ / the text / /
will be displayed in italics), while keeping the underlying wiki
syntax (I tried to do the same in syntax editors geany and kate). It
would be great to be able to offer this in a web interface for text
editing.
Otherwise, you who're interested in ASCII art, maybe you know
Dita, if this is not the case I advise you to see what can be done
with:
http://ditaa.org/ditaa/
Reply
fgallaire says:
22/08/2011 at 1:22
Hello Eric,
Thank you for your message, txt2tags also very lucky to have you you!
Aurelio is always overflowing (just like all of us :-)), but
as you said a release is a lot of work at once, and even if I think
too that the more often the better, I understand that in no hurry too
^ ^.
RedNotebook is actually very innovative, I will discuss in a
future article, as Dita elsewhere.
Reply
BEAUFILS Bruno says:
15/09/2011 at 16:00
An overview of interesting especially when the goal is to focus on
the bottom first and then the form (so can not do easily with the
tools "to" LibreOffice). It's nice to not be alone in search of the
grail.
That said I have just a small remark.
Certainly txt2tags is richer than docutils and rest for formatting
(** bold **, / / italic / /, and __souligné__-crossed-) but suddenly
we lose semantics and we commend the form where it should not be any.
From this point of view, I like the approach Markdown (just to
this aspect at least): * emphasis * ** strong emphasis **, and
(although I would have preferred to "focus" and * * strong emphasis ).
Reply
___________________________________________________________________
Best
A. Mani
--
A. Mani
CU, ASL, CLC, AMS, CMS
http://www.logicamani.co.cc
_______________________________________________
Ilugd mailing list
Ilugd@lists.linux-delhi.org
http://frodo.hserus.net/mailman/listinfo/ilugd
