[thelist] Discussing XHTML and ROI with your "boss"

[Michael Kimsal]

Mark Howells wrote:
>>It seems you're saying you make one 'page', then grab that and process
>>it via
>>XSLT or reparse it to extract data.
>
>
> The idea is to contain one page containing your content ("data") and then
> control it's screen, print or audible layout with a style sheet, be it CSS
> or XSLT.
>
>
>>If so, why do that?  If I've two
>>massively distinct agents to target, I'd just use two separate templates
>>(with XHTML or whatever in them) for the specific user agents.
>>Input comes in, gets processed, data for the template is returned
>>to the processing engine, which then calls the appropriate template(s).
>>I don't
>>need XHTML for that, and it affords the ultimate in flexibility.
>
>
> So, in your scenario, how would you handle the 29 different types of web
> browser that have visited my website in the past year? Have 29 different
> templates containing the same content (but different mangled code), or one
> XHTML template that they can all parse perfectly?
>
>

But they can't all parse it perfectly.  I'm dealing with the realities
of today.  The realities are that *most* can display things more or less
the intended way, but we operate with the understanding that no two
browsers are going to behave exactly the same way all the time.  This
WILL be the case even when everyone supports 'standards' 100%, because
there'll still be differences from different vendors.



>>If someone else is trying to screen-scrape me, and can't easily parse stuff
>>because it's not XHTML, tough luck
>
>
> The problem with your argument there is that the "... someone else ..." who
> is "... trying to screen-scrape ..." you will be the browser of your
> visitor, which is trying to parse the non-standard code that you are using.
> It'll read your code, try and make the best attempt at displaying it against
> it's internal processing SGML "rule set" and then, probably, show a
> sub-standard result.
>

Again, you're assuming that someone's browser will digest ONLY *100%*
valid-to-the-spec-published-on-this-day-and-nothing-else code.  That's
not the case, and I wholeheartedly believe it never will be.



>
>>I don't see this 'retargetting' thing as that valid an argument for XHTML.
>
>
> I'm not sure what you mean by retargeting -- one of the primary reasons for
> using standardized mark up languages (XHTML, HTML, XML, whatever) is that
> when a new browser or browsing device comes along, then the content that
> you've already marked up is ready for the new device. By correctly using
> mark up languages code that comply with a standard, you're helping the
> browser manufacturers to develop browsers which function correctly.

The 'retargetting' comment was a reply to the reasoning that the content
in XHTML could be reparsed and represented to another agent for a
different purpose.  (I don't think that was what I actually quoted, though).