What features will your dream XSL (I)DE have?

[Soren Kuula]

Hi, everyone,

I'm beginning work on my Master's thesis in computer science ...

I will start out with a pretty advanced, not yet published algorithm for
computing quite accurate control flow graphs for XSL style sheets. I
will hopefully end up with a prototype for a developing environment for XSL.

Now there are plenty of those already, but it occurc to me that they all
really only support:

- editing aid
- run time debugging

I will be able to throw static analysis on top of that, so that the
input and output schema are included in analysis.

So, XSL ladies and gentlemen, what costs you the most time and cinfusion
when writing XSLs?

- Difficulty figuring out in what order, if at all, templates are
applied to input?

- Uncertainty whether your output is schema-valid (and, when you boss
says, it HAS to bem what do you do?)

- Uncertainty whether you have remembered considering every possible
"quirk" and rarely used element/attribute/structure in the input?

And what are the features you miss the most? There are many possibilities:

- Static validation (guarantee that your output is valid if input is) ?
Do your bosses ever demand that anyway? Do you, yourselves?

- Visualization of possible control flow (this-may-apply-this-o-graph,
but depleted according to input schema)?

- Refactoring tools?

- Some other things?

The more experienced developers usually develop some way of doing things
to get around the worst obstacles. But you may try to recall, then, what
was the worst thing in the beginning

Thank you for any response

Soren

 

[Joris Gillis]

Do your bosses ever demand that anyway?

Do there exist bosses that demand of their employies to actively use XML and XSLT? Sounds great... As a student, I consider XSLT and XML more like playing toys. What companies have these policies?

- Visualization of possible control flow (this-may-apply-this-o-graph,
but depleted according to input schema)?

Should be done with SVG (computed with XLST of course


regards,

 

[anonymous]

Tempore 11:32:34, die Thursday 03 February 2005 AD, hinc in foro
{comp.text.xml} scripsit Soren Kuula


Do there exist bosses that demand of their employies to actively use XML
and XSLT? Sounds great... As a student, I consider XSLT and XML more
like playing toys. What companies have these policies?


Should be done with SVG (computed with XLST of course


regards,

You're in for a surprise. Almost all our projects in the last 18 months
for customers have included a need for xml, xsd, xsl.

So the tools aren't there yet? Tough luck!

 

[Joris Gillis]

You're in for a surprise. Almost all our projects in the last 18 months
for customers have included a need for xml, xsd, xsl.

You mean you are actually getting paid to mess around with xml and xsl? What company do you work for?

 

[Soren Kuula]

Tempore 17:35:23, die Thursday 03 February 2005 AD, hinc in foro


You mean you are actually getting paid to mess around with xml and
xsl? What company do you work for?

Lots do that.

So what will it take, in your opinion, to turn XML, XSL into something
serious then? Faster development turnaround?

Good ideas on how?

Soren

 

[anonymous]

Tempore 17:35:23, die Thursday 03 February 2005 AD, hinc in foro


You mean you are actually getting paid to mess around with xml and xsl?
What company do you work for?

Sure am. Most of what we do amounts to business functions. Their results
have to a) go on the web and b) go to 3rd parties using web services.
So, naturally, we use xml. This way we have only one set of data and
XSLT allows us to massage that set any way we want.

If you think this is exotic or very modern, think again.

Still, the tools are lacking severely, but they will (read: absolutely
have to) get there.

When Java first came out we used VI or Notepad, right?

We have three major government customers (Customs & Excise being one of
them), they demand the use of web services that have xml as input and
output.

Not being on top of this technology kind of puts one at the level of
PL/1, APL, COBOL developers. Those are, of course, just languages, xml
is something different, but you get the point..

 

[Joris Gillis]

Sure am. Most of what we do amounts to business functions. Their results
have to a) go on the web and b) go to 3rd parties using web services.
So, naturally, we use xml. This way we have only one set of data and
XSLT allows us to massage that set any way we want.

If you think this is exotic or very modern, think again.

What a revelation... When I first learned XML (about 2 years), I expected to see it being used everywhere, but I saw it nowhere. Apart from these mailing lists/newgroups I've never encountered a single person (not even an IT-guy) who knows XML. I began to think all companies were obstinately refusing to use any W3C recommendation. Good news to hear I was wrong. And I do perceive this as exotic (not the XML itself, but it's massive application you're implying.)

Still, the tools are lacking severely, but they will (read: absolutely
have to) get there.

I'm still using Notepad ++ to author xml and xslt, because, as I said, I don't use it on a professional basis. I don't feel like spending money on those visually bloated WYSIWYG xslt editors I've encountered so far. Possibly the really useful tools have yet to come...


regards,

 

[anonymous]

Tempore 20:39:10, die Thursday 03 February 2005 AD, hinc in foro


What a revelation... When I first learned XML (about 2 years), I
expected to see it being used everywhere, but I saw it nowhere. Apart
from these mailing lists/newgroups I've never encountered a single
person (not even an IT-guy) who knows XML. I began to think all
companies were obstinately refusing to use any W3C recommendation. Good
news to hear I was wrong. And I do perceive this as exotic (not the XML
itself, but it's massive application you're implying.)



I'm still using Notepad ++ to author xml and xslt, because, as I said, I
don't use it on a professional basis. I don't feel like spending money
on those visually bloated WYSIWYG xslt editors I've encountered so far.
Possibly the really useful tools have yet to come...


regards,

Well, since I use Linux, XMLSpy is closed land for me. Instead, I use
oxygenxml, it is 15% of XMLSpy's price and gives me what I need. It is
very cheap, I think. Java based, runs on anything you can power on.

 

[Dimitre Novatchev]

I'm still using Notepad ++ to author xml and xslt, because, as I said, I
don't use it on a professional basis. I don't feel like spending money on
those visually bloated WYSIWYG xslt editors I've encountered so far.
Possibly the really useful tools have yet to come...

You cannot imagine how far from the reality this is...


I have been using Xselerator for almost 3 years -- it is a great XSLT 1.0
IDE (great text editor with static and dynamic intellisense, support for
almost infinite set of different transformation engines, extremely
well-thought User Interface, native MSXML3/4 debugger)

Now we have Oxygen and Stylus studio with support for XSLT 2.0, the XML
editor in Visual Studio 2005 is great, ... and probably the list is quite
long.

I think the OP is a little bit late as the new XPath 2.0 data model and XSLT
2.0 have as their foundation strong and powerful datatyping. Functional
programming and higher-order functions in XSLT are also almost 3 years old.

So, in order not to be like the jokes series chukche, who wants to become a
writer and doesn't think it's necessary to read any significant literary
works to become a writer, one simply must read, read and read, before
deciding to make the next "great achievement". Probably the tuitors should
also read.

Just start with Michael Kay's last two books.

More seriously, I doubt that the topics listed in by the OP can really be
significant goals even for a Master's degree.

As for the original question, most dream features are announced after having
been implemented -- so watch this space.

Cheers,
Dimitre Novatche.