james_b
1/27/2006 10:35:00 PM
Henrik Martensson wrote:
...
>
>>I'm pretty sure Tim Berners-Lee,
>>Marc Andreessen, etc. knew about SGML, and I do not believe that HTML
>>ever had PIs.
>
...
>
> I can't prove that the people who wrote the first web servers did not
> know about PIs, but I think it is likely. If they had known, what
> possible reason could they have had for deliberately doing something
> that was not SGML compliant? (Browser wars and vendor lock in didn't
> become major issues until later.)
>
Tim Berners-Lee arguably wrote the first Web server (if Wikipedia is to
be believed). I'm doubtful that it had anything to do with processing
instructions; Web servers generally don't worry if the pages they serve
up are SGML compliant or otherwise; that tends to be left to the client.
The issues with funky page-generation syntax, at least for the Ruby Web
tools I've used, arise prior to interaction with any Web server.
There seem to be (again, generally speaking) two camps: Those who want
to treat the page templating or page generation source as a consistent
set of a single markup language (e.g. SGML or XML), and those who want a
special syntax such that an editor or other tool might readily
distinguish between document markup and programming language markup.
I'm guessing the latter describes the PHP/ASP/Erb path. Meanwhile,
Nitro, Amrita, Sean Russell's xml-tmpl, some others I can't recall,
allow for the use of either XML elements or processing instructions or
both.
I tend to prefer the PI-only approach, as it is (for me) less intrusive
for various XML tools such as tidy, and seems to make it easier to check
for certain page validation errors prior to full rendering (though at
some point that needs to be checked as well). Luckily, Erb lends itself
to some basic hacking to allow the use of PI syntax.
But I can see why people might want a syntax that was orthogonal to any
particular output format; Erb need not only be used to create Web pages.
It can generate Ruby or postscript or whatever, and a PI syntax would
mean little then.
--
James Britt
"Blanket statements are over-rated"