Hi Alan, On Wed, 14 Jan 2015 10:55:50 -0700, Alan BRASLAU wrote:

\enablemode[export] \startmode[export] \setupbackend [export=yes, xhtml=yes, css=export-example.css] \setupexport [hyphen=yes, width=60em] \stopmode \starttext \startquotation \input darwin \stopquotation \stoptext

LuaTeX, Version beta-0.79.2 ConTeXt ver: 2014.12.30 09:48 MKIV beta fmt: 2014.12.30 Your example doesn't work here. After 800 lines of overfull hboxes I get tex error > error on line 19 in file darwin.tex: ! TeX capacity exceeded, sorry [text input levels=127] l.19 \input darwin 9 10 \setupexport 11 [hyphen=yes, 12 width=60em] 13 14 \stopmode 15 16 \starttext 17 18 \startquotation 19 >> \input darwin 20 \stopquotation 21 22 \stoptext Please advise. Best wishes Idris -- Idris Samawi Hamid Professor of Philosophy Colorado State University Fort Collins, CO 80523

On Wed, 14 Jan 2015 12:58:35 -0700, Idris Samawi Hamid ادريس سماوي حامد wrote:

It's one of those weird things: Alan sent me the same file as a text attachment and it ran just fine. But the text I copied from his first email did not. Doing a quick compare of the two files in Notepad++ indicated that the two files are identical, but there must be some unseen spurious character caused by the email that context doesn't like..

No, this is totally off-track. The problem was that silly me named the test file darwin.tex, while there was an "\input darwin" in the test file. My apologies for the noise. Best wishes Idris -- Idris Samawi Hamid Professor of Philosophy Colorado State University Fort Collins, CO 80523

Hi Pablo, On Wed, 14 Jan 2015 13:15:38 -0700, Pablo Rodriguez wrote:

...

There must be some way to get this into Word...

pandoc may be the way.

I say “may be” instead of “is” because pandoc ignores some attributes when converting from XHTML

(http://pandoc.org/try/?text=%3Cp%3EThis+is+sample+with+a+foreign+word%3A+%3Cem+lang%3D%22de%22%3EFremdsprache%3C%2Fem%3E.%3C%2Fp%3E&from=html&to=markdown).

I was just about to ask about how pandoc handles xhtml. Some questions for you: Context produces three relevant files: darwin-xml-div.xhtml darwin-xml-tag.xhtml darwin-xml-raw.xml 1. Which one of these three files is the one we want to convert to docx? 2. I modified Alan's test file [same preamble]: ==== \starttext \startquotation \input darwin \bf \input darwin \stopquotation \stoptext ==== darwin-xml-div.xhtml and darwin-xml-tag.xhtml show up in the browser, but the bold does not. In Opera 12.17, darwin-xml-raw.xml gives a syntax error ==== XML parsing failed: syntax error (Line: 17, Character: 0) ==== But the "Reparse document as HTML" does work. In each of the three cases, there is no bold effect at all. What is needed to get the typography info transmitted? 3. My assumption is this: If I can get the xml/xhtml file looking right in the browser, I should be able to build a working docx file via pandoc. Thanks for any and all advice! and Best wishes Idris -- Idris Samawi Hamid Professor of Philosophy Colorado State University Fort Collins, CO 80523

On Wed, 14 Jan 2015, Idris Samawi Hamid ادريس سماوي حامد wrote:

==== \starttext

\startquotation \input darwin

\bf \input darwin \stopquotation

\stoptext ====

darwin-xml-div.xhtml and darwin-xml-tag.xhtml show up in the browser, but the bold does not.

What is needed to get the typography info transmitted?

Structural elements are exported; raw typographical commands are not. So, you would need: \definehighlight[important][style=bold] \starttext \important{This is important} \stoptext Aditya

On 01/14/2015 09:29 PM, Idris Samawi Hamid ادريس سماوي حامد wrote:

[...] I was just about to ask about how pandoc handles xhtml. Some questions for you:

Hi Idris,

Context produces three relevant files:

darwin-xml-div.xhtml darwin-xml-tag.xhtml darwin-xml-raw.xml

1. Which one of these three files is the one we want to convert to docx?

Only *-div.xhtml are (X)HTML files.

2. I modified Alan's test file [same preamble]:

==== \starttext

\startquotation \input darwin

\bf \input darwin \stopquotation

\stoptext ====

darwin-xml-div.xhtml and darwin-xml-tag.xhtml show up in the browser, but the bold does not.

Well, I cannot see any bold in "<div class="break"><!--empty--></div>". It seems to be how ConTeXt handles the blank line and the \bf command. BTW, the quotation environment is not translated as blockquote and paragraphs lack their <p> tags.

In Opera 12.17, darwin-xml-raw.xml gives a syntax error

==== XML parsing failed: syntax error (Line: 17, Character: 0) ====

It seems weird to me (that should be the error in Opera), that a <break /> element is placed outside the <document> element.

But the "Reparse document as HTML" does work.

I guess this format won’t be understood by pandoc (unless you write an specific reader for it). -raw.xml is not what you need.

In each of the three cases, there is no bold effect at all.

What is needed to get the typography info transmitted?

I have no experience with xhtml export in ConTeXt. This is beyond my knowledge. Sorry.

3. My assumption is this: If I can get the xml/xhtml file looking right in the browser, I should be able to build a working docx file via pandoc.

If it doesn’t look good in the browser, you won’t get it in pandoc. But it might be that you get it right in the browser and not in pandoc. It depends how ConTeXt outputs the XHTML. I hope it helps, Pablo -- http://www.ousia.tk

On Wed, 14 Jan 2015 15:22:39 -0700 Idris Samawi Hamid ادريس سماوي حامد wrote:

...

BTW, the quotation environment is not translated as blockquote and paragraphs lack their <p> tags.

Hmm, perhaps a showstopper.

Hans has often promoted the use of \startparagraph \stopparagraph for the particular reason that rigorous structure can thus be correctly exported. I still have a hard time following this in practice, as I find that paragraphs of text separated by blank lines to be more readable, although I do try to use \startxxx\stopxxx forms as much as possible. Example: \startitemize \startitem item one \stopitem \startitem item two \stopitem \stopitemize (although can someone indicate how to replace \sym{} ?) \startdigression Curiously, I sometimes work with coauthors who only know Word as their text editor. A work flow that we share is to edit ConTeXt source files that are easily editable by them. This means paragraphs without any line breaks, and the maximum use of UTF8 characters (no \alpha nor \int for example). I have trained them to understand math versus text mode and I place context commands on separate lines when possible. Since they still have a hard time understanding the significance (and the non-significance) of blank lines, I try to end all paragraphs with \par. I imagine that \startparagraph ...\stopparagraph would be a better practice in this workflow. \stopdigression The suggestion of using \definehighlight, etc. is good practice, too. Alan P.S. Mailers, certain text editors, and systems (MacOS for example) often add hidden characters or change the encoding of files. Many text editors hide these differences, so one must be careful. I understand, though, that the problem was that of recursion with an problematic choice of filename.

Am 15.01.2015 um 15:43 schrieb Idris Samawi Hamid ادريس سماوي حامد :

On Thu, 15 Jan 2015 01:01:37 -0700, Alan BRASLAU wrote:

...

The suggestion of using \definehighlight, etc. is good practice, too.

But to be consistent we would have to do this:

\definehighlight[important][style=bold]

\definestartstop[important][style=bold] Wolfgang

Hi Marcin, On Tue, 13 Jan 2015 14:34:21 -0700, Marcin Borkowski wrote:

On 2015-01-13, at 20:59, Idris Samawi Hamid ادريس سماوي حامد wrote:

...

Dear gang,

I'm currently writing a paper with very basic typographical needs: blockquotes, emphasis, bibliography, perhaps a graphic or two. The publisher wants the thing in Word, naturally. The last time I did something like this I set up a markdown document and just exported it to both docx and to context. For simple documents this is at least workable, but I'd prefer to write in context, not markdown.

While I /do/ understand you (it's the same with me, only that I'm more comfortable with LaTeX), have you considered Org-mode? It's similar to markdown (though better IMHO), but it comes with great support in Emacs. And you get export to odt "for free". (Also to LaTeX, though unfortunately not to ConTeXt; OTOH, writing a ConTeXt exporter should be fairly easy, and a simple exporter could be done in a few days - the framework for writing exporters to different formats is very well done. Incidentally, there's also a Markdown exporter, so you could probably get to ConTeXt via markdown).

Thank you, Marcin, for introducing me to Org-mode. OTOH, I decided some years ago that the Emacs lifestyle is a bit too much for me :-) Best wishes Idris -- Idris Samawi Hamid Professor of Philosophy Colorado State University Fort Collins, CO 80523

Hi Pablo, Thanks for your detailed reply: Somehow I missed it till just now. Comments: On Wed, 14 Jan 2015 12:58:11 -0700, Pablo Rodriguez wrote:

On 01/13/2015 08:59 PM, Idris Samawi Hamid ادريس سماوي حامد wrote:

...

Dear gang,

I'm currently writing a paper with very basic typographical needs: blockquotes, emphasis, bibliography, perhaps a graphic or two. The publisher wants the thing in Word, naturally. The last time I did something like this I set up a markdown document and just exported it to both docx and to context. For simple documents this is at least workable, but I'd prefer to write in context, not markdown.

Hi Idris,

I think it is better to use a markdown source with pandoc. It is the closest model to its native format.

I have written a small book from a markdown source. The PDF document was generated by ConTeXt, previously converted with pandoc.

I think this is the wrong way to go. Mainly because conversion to ConTeXt doesn’t rettain language information and special attributes. I had to edit the ConTeXt source to add the missing features.

Similar experience here; see below. But even the conversion to epub had issues, especially in getting the front matter to come out right.

[pandoc’s internal document format only allows attributes for certain elements, not for all of them. A workaround is to wrap then in divs an spans.]

In my opinion, the right way to go is to convert the markdown source to HTML and to parse it with an environment. Otherwise, updating the source is not straightforward (either you have two sources: markdown and ConTeXt).

What do you mean by 'parse it with an environment'? Could you give an example?

Why do you prefer to avoid markdown as a source document format?

I did a small-book project last year with markdown as source, with pdf and epub output. 95% or so was good but towards the end I had no choice but to decouple the context and epub files (so I had to deal with two sources). For that project I would say that the markdown/pandoc workflow was a qualified success. OTOH I was not the author of that small book. As a writer I am more used to writing in ConTeXt than in markdown; the flow of ideas is less disturbed that way. Closely related, the current project has a lot of bibliographical references and I don't want to manually write a bibliography in markdown.

...

Now it's been a while since I've experimented with ConTeXt and epub. From lightly perusing the list it seems that there has been considerable movement on this front. So my question: Has anyone here succeeded in the following workflow:

context -> xml -> docx

[perhaps via pandoc]?

I don’t have any samples of this workflow, since I never worked this way.

It seems from this round of experimentation that the pain of context->docx via xml and pandoc will be more pain than it's worth. At some point I suspect this sort of workflow will get sorted out for ConTeXt, especially since it is somewhat unavoidable for many users.

pandoc doesn’t even have a ConTeXt reader (https://github.com/jgm/pandoc/issues/1423). You will need a way to generate a document in an XML format that pandoc understands.

XHTML is an option. And I guess that would be easy to generate with ConTeXt (I have never done it).

Just in case it helps,

Indeed, your comments are always helpful, Pablo: Thanks again!! Best wishes Idris -- Idris Samawi Hamid Professor of Philosophy Colorado State University Fort Collins, CO 80523

On Sat, 17 Jan 2015 13:32:07 -0700, Pablo Rodriguez wrote:

On 01/17/2015 03:45 PM, Idris Samawi Hamid ادريس سماوي حامد wrote:

...

Similar experience here; see below. But even the conversion to epub had issues, especially in getting the front matter to come out right.

Sorry, what was wrong in the conversion to ePub?

Just getting a simple copyright page prior to the TOC page was a pain. I ended up having to patch the xml manually (IIRC you helped me in that thread...)

Unless you number your headings manually, you won’t get numbers in them.

As far as I know, automatic numbering is impossible in ePub 2 (there seem to be no counters available).

Is there any other feature you miss when converting to ePub.

As mentioned above, pandoc seems to have issues in getting the correct flow for Title-Page=>Copyright-Page=>TOC=>Main-Text. Transmitting this info to ConTeXt was also problematic. That's why I finally had to decouple things at the end of the process.

...

...

In my opinion, the right way to go is to convert the markdown source to HTML and to parse it with an environment. Otherwise, updating the source is not straightforward (either you have two sources: markdown and ConTeXt).

What do you mean by 'parse it with an environment'? Could you give an example?

Well, the ConTeXt Suite has at least one: tex/texmf-context/tex/context/base/x-html.mkiv.

Environments are the standard way that ConTeXt has to directly typeset XML files (http://www.pragma-ade.com/general/manuals/xml-mkiv.pdf).

Sorry, but I cannot send a more elaborate sample right now.

I understand now, thanks!

...

...

Why do you prefer to avoid markdown as a source document format?

I did a small-book project last year with markdown as source, with pdf and epub output. 95% or so was good but towards the end I had no choice but to decouple the context and epub files (so I had to deal with two sources). For that project I would say that the markdown/pandoc workflow was a qualified success.

Well, two sources is suboptimal at best.

Indeed!

...

OTOH I was not the author of that small book. As a writer I am more used to writing in ConTeXt than in markdown; the flow of ideas is less disturbed that way. Closely related, the current project has a lot of bibliographical references and I don't want to manually write a bibliography in markdown.

pandoc itself seems to be able to deal with bibliographies (I have never used them myself [either in pandoc, or in TeX]).

Perhaps I'll experiment, but *after* this article is finished :-) Thanks and best wishes Idris -- Idris Samawi Hamid, Professor Department of Philosophy Colorado State University Fort Collins, CO 80523