Steve Fosdick fozzy@pelvoux.demon.co.uk writes:
[...] What I have more trouble understanding is how XML differs from SGML.
XML is SGML-- rather than HTML++. It's just a bit simplified and easier to parse, I believe. The stylesheeting for XML (XSLT) is also supposedly easier to implement than for SGML (DSSSL), but that seems to be a matter of infrequent argument, as XSLT is Java-orientated while DSSSL is lisp-family-focused.
On 03-May-01 MJ Ray wrote:
Steve Fosdick fozzy@pelvoux.demon.co.uk writes:
[...] What I have more trouble understanding is how XML differs from SGML.
XML is SGML-- rather than HTML++. It's just a bit simplified and easier to parse, I believe. The stylesheeting for XML (XSLT) is also supposedly easier to implement than for SGML (DSSSL), but that seems to be a matter of infrequent argument, as XSLT is Java-orientated while DSSSL is lisp-family-focused. -- MJR
I agree with the statement that XML is SGML but easier to pass. SGML is a standardised version of GML which was a well done solution to the problem of how do we write a good type setting system.
What IBM did was it went out to speak to type setters and find out about the way type setting was done (specificly french type setters, please correct me if you know better) they then simplified the short hand layoutr system.
Essentially content providers labeld their content with <TAGS> and then the documents where passed on to page layout guys who added yet more tags, and finnally out to the guys wqwho made out the pages using the tags to build the page almost like a computer.
This waqs of cause being human full of exceptions and non standard bits, that IBM stripped out. They did not go so far as to lose all of this after all bytes costed in thoughs days. So SGML standardised this legacy system, which was a bit complex.
XML was just an attempt to rationalise this proses for the internet as no one used SGML for the internet as SGML people initially thought they would think thier skills would be good so they gave it another go and also their was a lot of little bits to SGML that could be lost with no hastle to any one and increase the speeed of the software and reduce the time taken to learn your way arround.
I would disagree though that XSL is like java, rather more like prolog meets a procedural language with SQL like libarys. SGML for me became instantly dated and of nothing more than historical interest after I had a look at it. as for DSSL i have not a clue.
Owen Synge
oms101@freeuk.com wrote:
On 03-May-01 MJ Ray wrote:
Steve Fosdick fozzy@pelvoux.demon.co.uk writes:
[...] What I have more trouble understanding is how XML differs from SGML.
XML is SGML-- rather than HTML++. It's just a bit simplified and easier to parse, I believe. The stylesheeting for XML (XSLT) is also supposedly easier to implement than for SGML (DSSSL), but that seems to be a matter of infrequent argument, as XSLT is Java-orientated while DSSSL is lisp-family-focused. -- MJR
I agree with the statement that XML is SGML but easier to pass. SGML is a standardised version of GML which was a well done solution to the problem of how do we write a good type setting system.
on a content/markup note.... If anybody needs any help with Latex, let me know, I'm now finishing off my Phd thesis, and believe me, Latex is definatly worth learning.....
nothing to do with xml/xslt/sgml, but it seemed a good point to drop it in ;) Sz
On Fri, 4 May 2001, Neill Newman wrote:
on a content/markup note.... If anybody needs any help with Latex, let me know, I'm now finishing off my Phd thesis, and believe me, Latex is definatly worth learning.....
What XML does is allow the database developer to define the _content_ of what will eventually become a web page using application-oriented tags and without having to worry about the page's appearance. That's why we have MathML, SyncML, VoiceXML and all these other not-quite-agreed markups for particular applications. In principle (though I've never seen it done) XML could produce LaTeX so that we could get typeset output that looks better than what spews out of Netscape, which is definitely of the same kind of abysmal quality as Word.
This separation of content and appearance is precisely what Tim BL (name-dropping) was actually trying to do with HTML before Netscape jumped in and started adding appearance-related tags: the page defined content, the browser appearance. Oh, how I wish we could go back to those days so that I didn't now have to read pages filled with miniscule Arial text!
The advantage XML over (say) keyword-value pairs is that XML allows the developer to build up a tree-like structure that can be parsed and processed by a standard piece of software.
Anyway, Neill is using lyx, not pukka LaTeX -- urghhhh!
..Adrian (Neill's PhD supervisor and a LaTeX hacker ;-)
"Adrian F. Clark" wrote:
Anyway, Neill is using lyx, not pukka LaTeX -- urghhhh!
;p .. lyx produces nice latex ;)
..Adrian (Neill's PhD supervisor and a LaTeX hacker ;-)
LaTeX guru more like ;)..
Sz
MJ Ray wrote:
Steve Fosdick fozzy@pelvoux.demon.co.uk writes:
[...] What I have more trouble understanding is how XML differs from SGML.
XML is SGML-- rather than HTML++. It's just a bit simplified and easier to parse, I believe.
yep, sounds about right to me, I lent out my xml book so I can't look up exact difference right now ;(
The stylesheeting for XML (XSLT) is also supposedly easier to implement than for SGML (DSSSL), but that seems to be a matter of infrequent argument, as XSLT is Java-orientated while DSSSL is lisp-family-focused.
hhmmm I'm not sure about the 'java orientedness' of XSLT, as far as I understand it does collect data in a loose 'object' structure, but I wouldn't call it Java oriented by any stretch of the imagination ;)
Sz
-
Adrian F. Clark -
MJ Ray -
Neill Newman -
oms101＠freeuk.com