Just to make this crystal clear:
-log2html- is a user-written command from SSC which is a wrapper for
official Stata's -log html-. (-log html- is not documented.)
David's diagnosis looks spot on to me.
For the reasons he states, and others, my guess is there will be little
or no future development in -log2html-, but the other two authors
maintain splendid independence of thought and action.
Nick
[email protected]
David Elliott
Since replying earlier in this thread regarding -log2html- and an
out-of-thread conversation with Ashim I have done some more poking
around in -log2html- and the output of -log html- which is used to
translate smcl to html (which it does badly, I'm afraid)
-log2html- takes the output of the -log html- translation and attempts
to create a version that gives better indication of input, result,
error, and text colours per Stata's usual results window output.
However, it has to infer regarding error output and only puts the
error code itself, e.g. r(197);, in the error colour. If you
specified -di "{err:This should be red}" you are out of luck - it will
display as ordinary text. This is not the fault of -log2html-, it is
a consequence of -log html- dropping any unique tagging of {err} text.
It really is a shame that a better conversion using stylesheet
classes can't be accomplished. It would then be a simple matter to
have <span class="txt"> Domestic |</span><span class="res">
52</span> to do the display work. Since the translation works in line
mode and SMCL uses nested tags in a manner similar to HTML I don't
think it would be that difficult.
However, I still believe the best answer is still to work toward a
table-based display paradigm (as opposed to the current line-based)
for explicitly tabular data. Version 8 brought us a new graphics
engine, perhaps version 11 needs a new table engine?
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
