Re: Microsoft extensions to XSL (was RE: how to call Javascript function in .xsl file) function in .xsl file)

[Tyler Baker <tyler@xxxxxxxxxxx>]

Didier PH Martin wrote:

> Hi Tyler,
>
> Simple question. Did you created something with DCOM?
>
> We are working with both since 4 years now (COM at that time and a beginning
> CORBA). I don't think that CORBA is superior like you say, only that CORBA
> and DCOM are different having each one strength and weakness. A weakness
> example of DCOM. Actual security is NT only. Actual strength, you can bypass
> it by replacing the marshalling. CORBA strength: object inheritance, CORBA
> weakness object cannot have multiple personalities accessed through a query.

Not true.  Typing is done via CORBA interfaces.  What actual implementation you
have on the other side is transparent to the caller.

>
> So I personally think that both object middleware should take the good
> things of the other. One thing remains, both are independent of languages
> and take a architectural approach to design, which is a good thing, remember
> the facade pattern (Gamma and al.) with interfaces you can have more than
> one implementation sharing the same interface. This is what we have with
> script engines.

This is true, but the experience I have had with COM is somewhat painful.  The
worst thing about it is there is no real decent Java implementation out there
and that DCOM unfortunately sticks you with Windows and the Windows registry.  I
know there are so-called COM implementations for other platforms, but I have
never heard of any success story with these products.  COM is more or less a
service protocol.  You can design the same thing easily and transparently with
CORBA to mimic this pattern.

> So, come on, beyond the wholly war there is serious thinking. Each time,
> someone is doing statements like that is like saying the earth is flat
> without taking a boat and trying to see if that is true. Don't mix business
> practice of a horde of barbarous people with technology which could be
> evaluated as objectively as we can. Is it how we made progress in sciences?

You have a good point here, but my feelings toward COM have little to do with
Microsoft's business practices.  My experience (though not as great as yours by
any stretch of the means) is that it is an ugly mess.

> So the point here is. Something is missing, a tag definition for scripts. It
> is possible to have more than one script. the suggestion that we have a
> standard script as default makes sense. However to be able to choose one is
> a plus. Actually, any XSL implementation able to use these DCOM (which are
> in fact C++ Vtabl when used in process - a certain progress to "C" dll
> doesn't it) object could in fact support more than one language and this has
> nothing to do with Microsoft. Perl is not made nor owned by Microsoft, so is
> Python. If people use DCOM it is not to support Microsoft, it is a practical
> solution with good technical points and weakness . Just ask Perl or Python
> implementers. If Netscape proposed a good "in process" interface mechanism,
> we could see also different implementations. But maybe the best location for
> such feature would be on the server not the client.

What you are looking for here is some sort of post-processing solution.  A
client I have is using JavaScript with XSL just fine (I don't mean JavaScript is
being used to transform the source tree only that it is in the output).

> About XML mutation - don't forget that XML is a language to design
> languages, so wait a couple years and count how many languages are derived
> from XML. History will speak.

Well, I have no problem with that.

> About XSL not being a language - What is your definition of a language then.
> Come on, don't take us for fools. What about a transformation language?

I was speaking of programming languages.  XSL is not a programming language, it
is a transformation language as you have noted.

> If XSL should not include other languages. It is maybe controversial from a
> philosophical point of view but quite particle in fact. A strong point of
> any XML derived language is that you can include more complex constructs so
> why not include a script? For religious person that should not include a
> script in their XSL, let them live as they wish and respect their
> principles. But for others with real practical stuff to do, why not let them
> include script in an _ordely_ manner. Does diversity sound like something to
> you? I am wrong to say that one goal of XML is diversity of languages
> derived form it? If it is not, we do everything to let that happen.

I am arguing that core XML as it now stands should not mutate a lot.  What you
actually do with XML in your application is proprietary anyways.  In an app I am
working on, I use it for everything for all sorts of things.

> Is it worse to let some include scripts in a transformation language than
> everybody define their own language with XML than letting ?

I think the worst thing is to try and use XSL for something that it was not
suitable for.  XSL is not the end-all solution to everything.

> Is it because Microsoft did it that you suffer so much? so, why not define a
> standard not owned by Microsoft. XSL has to not say how to implement these
> script language, just how to invoke them.

No, it is because they referred to these tags in their documentation as if they
were formal parts of XSL as it now stands.  If anyone else did this, they would
not get away with it, but Microsoft has successfully gotten away with
misinforming people over and over for their own benefit in the past.  If
Microsoft did not have such a long history of serious ethical problems, then I
would just assume their engineers made a mistake and that this act was not
deliberate.

> Finally, to change from one dictator to an other is not necessarily freedom.
> Java is owned by Sun and the whole platform has not the same status as Linux
> or Appache.

Of course not.  But isn't XSL supposed to be some sort of public draft?  Why
should a company like Microsoft or SUN for that matter have undue control over
bending these standards for their own needs.

> And about Netscape, maybe they didn't got enough money which is sad, to
> improve their architecture to have it more "component" aware and be less
> "big application". their open process is not as open as Linux or Appache so,
> I guess that Navigator won't see the same evolution as these guys got.
> Should I remind you that these project just took what they found is either
> useful or good on the other platform and included it. This is evolution.

This is an XSL list, so lets not get started here on politics.  I really wish
that MS had not tried to pull a fast one, so we could all be talking about XSL
and not a deliberate attempt to sabotage a public spec for their own benefit.

> Wholly war, do not produce good minds, just dead bodies, you know.

Well, dead bodies are also produced from letting others walk all over you.
There is a long list of companies that have been killed as the direct result of
anticompetitive business tactics from large ISV's.  If you are a small fry like
the majority of developers are, then you better stand up to sly tactics by large
ISV's or else not even bother trying to start a software business.  Bullies and
criminals pick on the weak and apathetic.  When faced against tough odds they
rarely play by the rules and instead result to cowardly tactics which often
times are legal, but highly unethical.  As developers, we are not judges in the
legal sense but we are judges in identifying who is playing by the rules and who
is not.  Microsoft unabashedly with this XSL issue does not seem to want to play
by the rules.

> So, if someone from the W3 XSL group is participating to this list (Is W3
> open? R: is membership free? are discussion openned? Is IETF opened Yes, is
> particiaption free: yes, so we have to deal throught the elite it seems,
> like in politics), can we influence the process? could we prevent that
> Microsoft give us what we need and be locked again solely on their platform,
> only for religious reasons... silly.

Just working for Microsoft, or SUN, or any big name software company does not
mean you are elite.  The brightest minds in the US at least when graduating from
graduate school tend to get involved in startups these days rather than a rigid
corporate structure.  Most people do a little of both in this industry, but to
say that just because you work for some big name software company that you
deserve to dictate process to everyone else should expect resistance to this
philosophy unless all software developers are geeky, wimps who cannot stand up
for themselves.

> By the way tyler, the Javalobby stuff is quite impressive and well done.
> However, most of us has to do more complex sites, like for instance provide
> e-commerce and link to mainframes (yes, these prehistoric beast are still

In this case I would say that XSL may not be the right solution here or else you
will need to mix and match other technologies to get things to work the way you
want.

> alive), link to relational database, etc. XSL is great in a way that it
> provides a good transformation language, javascript is not so efficient for
> that task. But XSL is not able to link to database and build a document from

Using server-side Javascript has always been something I have been skeptical
about.  Most people I know seem to prefer the PERL solution or another popular
solution in Java servlets if Java happens to be something you know well.

> data stored in these beast. And if a standard can help us to do that, we
> won't get locked into any religion either called Microsoft or Java. I hoped
> that people from W3 beyond all religions heard this request from people
> having to _deliver_ stuff.

I totally agree.  But I don't want to get locked into a standard that makes
references to DLL Hell either.

Tyler


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list