On Jan 24, 2013, at 1:42 AM, Rainer M Krug wrote:

> Hash: SHA1
> On 24/01/13 01:02, Jerry wrote:
>> On Jan 23, 2013, at 1:02 AM, Rainer M Krug wrote:
>> On 22/01/13 23:37, Jerry wrote:
>>>>> On Jan 22, 2013, at 1:38 AM, Rainer M Krug wrote:
>>>>>>> Also, I played with pandoc a while back and its conversion of a very 
>>>>>>> simple LyX file 
>>>>>>> to Word was not perfect. I went LyX -> LaTeX -> docx. I did not 
>>>>>>> investigate setting 
>>>>>>> options very much. I recall that equations made the trip and were 
>>>>>>> editable in Word
>>>>>>> 2011 (Mac) built-in equation editor, but equation numbers were lost. 
>>>>>>> Today's effort,
>>>>>>> also on a simple but different document, LyX -> LaTeX -> docx, one 
>>>>>>> equation was not
>>>>>>> typeset (only the markup appeared), a .eps file was not found even 
>>>>>>> though it was
>>>>>>> present, and section labels were translated as literal text. Converting 
>>>>>>> to .odt was
>>>>>>> worse, and at least once caused it to crash when it was attempting to 
>>>>>>> repair what it
>>>>>>> thought was a damaged converted file. (But what _doesn't_ cause 
>>>>>>> LibreOffice to
>>>>>>> crash?)
>>>>>> I also tried to go the LyX -> LaTeX -> docx route, but the results were 
>>>>>> not as usable
>>>>>> as via xhtml. So I would suggest to try the route via xhtml.
>>>>> How did you get from XHTML -> docx? Pandoc (according to the first 
>>>>> paragraph of the
>>>>> user's manual) does not accept XHTML input files. Jerry
>> Well - using
>> pandoc -o newfile.docx newfile1.xhtml
>> works as expected and produces a docx file.
>> So I can only say it works. I only used the converters and format 
>> definitions as mentioned 
>> earlier.
>> Truy it:
>> simple lyx file, export to LyXhtml and use above command to convert to docx.
>> Cheers,
>> Rainer
>>> Thanks, Ranier. I have just now tried this on a simple document (two 
>>> equations, two sections,
>>> a footnote, a figure with caption, a greyed comment, and two cross 
>>> references (to an equation
>>> and to a section).
>>> As usual, the LyXHTML looks very good. But in the docx, the equations are 
>>> not rendered and 
>>> contain spurious text, the figure caption and figure name are separated 
>>> from the figure and
>>> are shown as normal text, spurious text is inserted relating to the image, 
>>> and all of the 
>>> cross-references move the cursor to the title rather than to where it 
>>> should move. The
>>> footnote is printed as ordinary text as is the greyed text also.
> Without doubt this is not perfect. As I said, I can't comment on equations 
> and I did not worry
> about the captions. Have you tried LyX -> eLyXer -> pandoc as well?
>>> pandoc -o newfile.odt newfile1.xhtml
>>> results in a similar result, but with completely nonfunctional cross 
>>> references and no
>>> spurious text relating to the figure image.
>>> I'm using OS X, LyX 2.0.5, Word 2011, LibreOffice, and pandoc 
> For my case (no crossreferences, no inseted references, but formating in the 
> text (italic for
> species names) and only some pictures which I deleted anyway to reduce the 
> size of the docx as I
> had to email it for further revisions it worked better and faster then any 
> other solution.
> Have you tried to go ia LaTeX and pandoc?

Yes. I have commented on this recently.

LyX -> LaTeX (plain) -> docx using pandoc:

Word (Mac 2011) complains "This file is corrupt and cannot be opened." It then 
offers to repair the file. Results are: inline equation translated and editable 
but with two minor errors (it did not recognize lim as a function and set it in 
italics)--this shows that equation translation _can_ work; standalone equation 
not translated, TeX code is displayed between two $ signs, and equation number 
is lost; image caption is lost, but the spurious word "image" is displayed 
nearby; labels (section, equation) are set as ordinary text within their 
respective domains; cross-references are displayed as ordinary text but with 
nonexistent links; comments are displayed as normal text but in a separate 
paragraph, thus splitting the paragraph in which it is contained into two 
paragraphs; and indexed terms are lost.

LyX -> LaTeX (plain) -> odt using pandoc:
Opened with LibreOffice. The file is reported as corrupt and the program offers 
to repair it. Results--similar to above for Word except: inline equation is 
very poorly typeset and is not editable using the built-in equation editor; 
instead of the picture that was supposed to be a figure, a box is displayed 
which reads "Read error"; the word "image" is not displayed; clicking on the 
footnote cross reference moves the cursor to the crossreference;

> I guess some manual fidelling will always be necessary, and for my document 
> the fidelling required
> after pandoc was considerably less as after any other approach.
> Please keep us posted about any progress.

I'm afraid there won't likely be any more progress as I have spent a lot of 
time on this already, plus I have exhausted most of the obvious options. OTOH, 
if one considers the combinatorics of combining various options, I'm sure there 
are more options than any one person has time to try.

Unless there is progress from other quarters in the meantime, I suppose when 
the occasion arises to make a conversion, I'll look into commercial options on 
Windows. As a Mac user, I can of course run Windows, but I don't relish the 
idea of buying another copy of Office, plus converter software for Windows. I 
don't mind paying for commercial software except Microsoft stuff which I find 
is generally of poor quality (in fact, I often prefer it) but the extra hassle 
of going this route, with still unknown results, is not appealing. I think I 
recall a Windows converter from LaTeX to Word where there was also offered a 
manual conversion of anything that their software did not convert, for an 
additional fee per megabyte over a certain level of file size. What a mess.

Thanks for your suggestions.

> Cheers,
> Rainer
>>> Jerry
>>>>>> You should try both converters (the build in and eLyXe), as they 
>>>>>> performed differently
>>>>>> on different objects. As my document did not contain any equations, I 
>>>>>> can not comment
>>>>>> on that.
>>>>>> I must say, that I also used pandoc to convert to odt and then converted 
>>>>>> the odt to
>>>>>> docx with OpenOffice (this was before LibreOffice...) and it also worked 
>>>>>> well.
>>>>>> I can only suggest to try the different paths out and see which works 
>>>>>> best, and then 
>>>>>> post your experiences here and add them to  
>>>>>> http://www.lyx.org/trac/ticket/6042 so
>>>>>> that there is some progress with pandoc support in LyX.
>>>>>> Cheers,
>>>>>> Rainer
>>>>>>> Jerry
> Version: GnuPG v1.4.11 (GNU/Linux)
> Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
> p5HaBDZEpcpPexLrUgwHCrWVp7jPHGGt9qP+InnyU5PTizsPwXCoh8l0HkDYWQ+n
> Kz2FhV+hUuWtlwR5gdV59nEPDnWy7auQUgFnZqy6JQ/rSdrBnfZ8EXi0yD250YtN
> xWaTnjSGn3Pa+dlL/crAGM+hm+bEqs23wzssQvZvunF9lx5LnnoitIwuggcdVyxo
> YNyN4VWj7vPUmRle7Qt1zhyZSn8Wb2TV4ZziKvtGa9couzXoq2tqUeIDpoCIDASR
> lXwEWY9sKb7uPybnnqUckAEE8TgQIFkNfpsWPnJRgiiS0fp5zd3FCUNsWCn2LHQ=
> =xga4

Reply via email to