Re: st: Creating HTML from SMCL log file including graphics - revised log2html.ado

[Phil Schumm <pschumm@uchicago.edu>]

On May 19, 2006, at 8:35 AM, John Novak wrote:
> It seems to me that the SMCL folks need to be talking to the XSLT  
> folks.  I would not even classify myself as a novice with respect  
> to XSLT, but from my very limited understanding it would seem to be  
> perfectly suited to handle tasks like translating code from a  
> structured markup language into viewable HTML.

XSLT would not be useful here, since SMCL is not XML.  What would be  
useful, if someone wanted to undertake it, would be to write an XML  
schema capable of representing all of the SMCL constructs.  Once this  
was done, one could then write a utility (perhaps in Mata, but  
probably in another language) capable of parsing SMCL and translating  
it into this hypothetical XML document format.  At that point, people  
could begin writing XSLT stylesheets to translate this into HTML,  
LaTeX, PDF, DocBook, OpenDocument, etc.  Moreover, one could  
programmatically manipulate the resulting XML document to do all  
sorts of things.

Of course the real question is, is this worth doing?  Consider three  
cases:

1) Using Stata to analyze and write papers.  While the world would be  
a better place if all journals accepted submissions in an open  
format, I suspect that many users are submitting to journals that ask  
for Word format.  For these poor souls, doing anything more with SMCL  
output would probably not be helpful.

2) Using Stata to prepare class materials, web-based instructional  
materials and simple reports, etc.  Users doing such things might  
benefit from more output flexibility and processing tools, however it  
may be that existing approaches such as those used by -log2html-, - 
grss-, -do2htm- are sufficient in most cases.

3) Using Stata as part of a larger system to programmatically  
generate reports and other types of documents.  This is where a  
comprehensive approach would be most useful, however users doing this  
have probably already put a lot of investment into established  
procedures, and so such an approach would have to be flexible and  
modular to be of value to them.

My only point is that I think a serious discussion of use cases  
should proceed the making of grand plans to write additional  
processors for Stata output.


-- Phil

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/