Subject: SAX is not XML (Key() Re: Saxon VS XT) From: Paul Tchistopolskii <paul@xxxxxxx> Date: Mon, 07 Aug 2000 19:19:49 -0700 |
----- Original Message ----- From: Thorbjørn Ravn Andersen > > > Not using key, is like having to use Perl > > > (or any other programming language) without being allowed > > > to use hash tables for lookup purposes. > > > > Poor C, ( and Pascal ) they had no build-in hashtable support. > > No. I am talking about not being allowed to use the technique at all. XSLT is using this technique already. There is some intersting twist, I'l write more later. > No. I just think that your reference to "most developers" is incorrect. If it make us closer to any solution let us agree that "most developers" will use key(). I don't care listening that I'm an idiot, if it get's some fair argumentation, because I'm concentrated on the technical problems but not on the style of presentation and style of discussion of those problems. I'l explain the particular problems with current key() semantics in a huge letter which I'm writing for last 5 hours. Thanks to all who participated in the thread. > > "XT has many limitations because it is not 100% conformant > > but conformant engines have no limitations because they are > > conformant" ( to me the only possible argumentation is key() > > and as I already wrote to Sebastian privately - I need at least > > 2 weeks to implement my model of processing his XML > > real-life data without key() ). > > Why not use the two weeks to implement key() in XT instead O:-) My next letter will have the answer for this also. The answer is shocking, as it was shocking for me to realize what have really happened 5 hours ago. The answer was in my letters, but I was not seeing it. > > Those lost souls who still think that storing massive volumes of > > data in the format of huge text files and then use key() hack to > > 'improve the speed of processing' - is reasonable usage of > > computers - I can not help them ;-) > > If that is your concern, why are you using XML in the first place ;-) I'm not using XML files for anything but tiny configuration files, simple web-pages, layouts and other stuff like this ( this is what XML syntax is good for ) I occasionaly made a mistake using XML syntax to express some templating language, but then I corrected my craziness with XSLScript, so - no more abusing XML syntax for something which could be better done in Yacc. Of course, I intensively use *SAX* - this means I'm using some subset of XML *model*. But because the subset I'm using is extremely generic, it should be called 'MinML', not 'XML'. PXSLSevlet takes SAX events out of SQL database, feeds XT with the stream of SAX events and then XT produces HTML. Where is 'XML file' here? Is there any need for that 'XML file' ? No need. Where is the 'XML model' with all those PI's e t,c ? No need. I almost not use most of XML macroprocessing constructions ( they are evil and could be easeily implemented in XSLT ). For example I'm using <X/> + <xsl:template match="X"> to render the word 'XSLScript' but not the entity &X; I'm trying to use the reasonable subset of any technology, when I use it, because I belive in common sense. Common sense says me that SQL is good for storting relational data. To me SQL is *not* "select bla-bla from dual". I don't understand how can I express this better. Perl, or 'java' or 'XML' is not a panacea. XML *syntax* is good for some limited number of applications. ( mostly 'dumb data' ). XML *model* is bloated with fake entities - which requires the bloated API's. But because SML-dev list have produced the really usefull subset of XML model and because XSLT produced really useful processing technology ( more in my huge letter ) - I'm taking those 'parts of XML' which I need - and that's it. I don't think storing Sebastian's data in plain text file is appropriate. But I understand why he does it. He is using the best free technology he has at the moment. In 10 years he'l get a better tools. ;-) ( OK - he already has something and will get some more tools much faster than in 10 years.) Rgds. Paul. XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
RE: Saxon VS XT, Thorbjørn Ravn Ander | Thread | Announce: PHP Class Wrapper for Sab, Bill Humphries |
RE: Excluding nodes, Matthew Bentley | Date | XSL Filtering, Bryce Wells |
Month |