Xsl

[Cowboy Bob]

It's like the difference between procedural languages and object orientation. If you were a COBOL programmer, then OO would seem completely alien. Likewise, until it "clicks", XSLT seems counter-intuitive.

My advice is to not try and force it to work in the way you think it should, and to experiment to see how it really does work. You'll get more satisfaction that way, and there'll be a lot less hair removal...

[DimPrawn]

AtW, surely the best approach is to write your own transformation language in assembly, targeting specific CPU's, and optimise it to the n'th degree.

[AtW]

XSLT - XSL + XPath is something like SQL for XML. There is indeed a need for such lightweight query and formatting language, that's why I picked it. However just because it deals mainly with sets (like SQL) it does not mean that it should not support variables, loops etc - all proper databases support such features in SQL. Now using custors is evil and I never did that because its possible to avoid it, either using full set operations or procedural elements.

Now XSL people say its Turing complete, in other words it can do like anything C/C++ do. They can kiss my arse - its the same as to say that in theory a person can walk around the world, and there are nuts who actually did that, but it does not mean that its good ffs.

Where is Haskell?

Where is Scheme?

Where are other dozens of weird languages creators of which proudly were saying how different they are and that its for different mindset. Well they can F\/CK Off with this mindset crap: there should be fking amazing advantages offered by change of mindset, this mind fk is not something that is productive and all they had to do is add a few procedural elements to make real life easy. Of course you can do it all different, the XSLT way, but it takes more time ffs, and time is money.

Sadly there is no alternative, and luckily I've got C# layer on top of XSL to enhance its functionality.

And oh - the fact that you can't have template that will put HEADER stuff like open TABLE tag, and another template that will put closed TABLE tag is bullcrap - they should validate FINAL document, not parts of it.

[DimPrawn]

If you have some spare time read this:

http://www.defmacro.org/ramblings/fp.html

It will explain why FP exists, why it is a good idea and why it is Turing complete.

HTH

[mcquiggd]

Alexei, would you mind your fuvking language, old boy? You've become terribly chav in recent times....

[AtW]

Look, its like this - this thing is also Turing complete:



It was made from Lego blocks, see here.

So, its Turing Complete just like XSLT, does it mean that by the virtue of being Turing Complete they are suddenly good designs or really useful in real life? No.

Turing completeness is as bull as suggestion to walk around the planet instead of using other means of transport. IT is really full of nutters who completely ignore things like usability and real life experience - that's why Microsoft kicked Unix butt, even though technically Unix is superior, turing complete etc etc ... but why would I give a fk if it would take me more time to do actual added value things?

Functional my arse - why they include IF, but no ELSE, even though they include CHOOSE, which is SWITCH by the other name? IF and ELSE is basically CHOOSE one and OTHERWISE, and yet they ignore IF/ELSE contract that is so familiar to millions of programmers. Why? They already have such functionality in place, why they design some language that is so fking alien without good reason. Is it not weird already to avoid ELSE for IFs?

The guys who designed it are clueless - the only reason they got ANY traction in it (and it is not used widely, now I see why), because they jumped onto XML bandwagon and got support from W3C.

[AtW]

Alexei, would you mind your fuvking language, old boy?

[Cowboy Bob]

The guys who designed it are clueless - the only reason they got ANY traction in it (and it is not used widely, now I see why), because they jumped onto XML bandwagon and got support from W3C.

I respectfully disagree. I find it quite elegant and VERY good at what it is designed for.

As for being used widely, in the last 3 years I've come across 3 companies who use it, one very extensively using FOP to generate PDFs. Also, you'll find that nearly all the Apache websites are generated from DocBook data and transformed into HTML. They also have templates to turn the DocBook data into other formats which is where it becomes really useful. I've also started to do similar things with my documentation. For example:-

This + this transformed through this = this

[bogeyman]

My advice is to not try and force it to work in the way you think it should, and to experiment to see how it really does work.

Sound advice there Cowboy.

I think AtW is suffering from the "if the only tool you know is a hammer, everything looks like a nail" syndrome.

I'm pretty sure XSLT was conceived by more able computer scientists than our own esteemed AtW. It works for me anyway. Not sure what he's doing to get himself in such a mess with it.

[Cowboy Bob]

I'm pretty sure XSLT was conceived by more able computer scientists than our own esteemed AtW.

I've actually worked with the guy who invented it - Dr Michael Kay - when he worked for Software AG. He's a VERY clever chap. Mind you, he was partly responsible for this monstrosity as well which I don't think I can ever forgive him for...