RE: Venting

Subject: RE: Venting
From: "Didier PH Martin" <martind@xxxxxxxxxxxxx>
Date: Thu, 4 Feb 1999 16:51:40 -0500
HI Guy,

Can I ask the DSSSL bods out there something, as I have no experience with

As I understand it, DSSSL expresses both transformation and flow objects
(equiv. formatting objects). Would DSSSL users appreciate the split of

It is also the stated goal of the XSL WG to achieve for XML a style
language at least as expressive as DSSSL and CSS... why should we have less
than that for XSL?

Let's talk a bit about dsssl. As we know this is an ISO standard. Some said
that it is not user friendly, some said it has to replaced because the ()))
stuff gives them headaches etc... all kinds of good reasons. W3 and its
members decided to create their own stuff and not respect nor let evolve
DSSSL. You know Guy, standard institution are like companies and there is
competition among them so ISO and W3 are not necessarily in the same camp,
and may that sometime can be perceived like two competing companies.
Nevertheless, W3 decided to compete against ISO instead of improving what's
there. For all kind of good reasons I am sure. But one for sure, HTML
difusion gave W3 incredible visibility and and a strong brand (When I say
brand, it is not that i say that W3 is a deposed trademark but that W3 name
is more important know and have influence power or power to create action,
in one word - it well known now)

DSSSL implementation is actually solely reduced to a single implementation
Jade created by James Clark. All known derived work is based on this engine
except for a Scheme engine named Bigloo but bigloo only implements the
template for DSSSL not the processing stuff. At a certain time there where
an other implementation created in Java but it died. Jade implementation
took completely the opposite direction than XSL implementations are taking
now. If you look at DSSSL specs, it is stated that DSSSL is at the same time
a transformation language and a style sheet language. All formatting objects
are well documented. this spec integrated most concept also behind XSL like
grove and property sets that themselves came from Hytime. The fact is that
the only implementation did only implement the formatting objects and not
the transformation part made that concretely speaking people could use only
formatting object but not the transformation part. James invented a clever
mechanism to transform using formatting object but this was not the
transformation part of the specs.

So what happened? totally the opposite of what's happening today with XSL.
XSL concrete implementations except one. started with transformation and to
be more specific with XML->HTML transformation in mind. Not a bad thing, of
course, because it is grounded in real commercial issues and market need.
Not 100% of the market have a HTML 4.0 compliant browser so imagine now
having a XSL compliant one!!!! Like I said earlier, we tend to forget
technology introduction cycle which takes always more time than we think it
takes and if we doubt, let's just sit down for a moment and think why this
list is not using MHTML mail even if is a 2 years old standard (to reassure
people - MHTML does not mean Microsoft HTML but Mail HTML :-). What we have
to learn from DSSSL story is that the real standard has been imposed by
concrete implementations. Just ask today people about DSSSL and you'll have
probably the linkage to Jade. So DSSSL is now synonymous to Jade. Like if
most implementations Microsoft included, includes only the transformation
part, XSL will be associated to the transformation part. There is an
historical difference, W3 brand create quite a lot of action and conviced at
least 5 to 10 people or companies to create a XSL implementation. ISO brand
didn't created so much activity. Now that W3 brand has motivated people to
invest time and money and that we have healthy competition and users having
liberty of choice. These producers and some erly adopters are asking W3 to
give them some chance and have their work be "standard" compliant. Like you
said W3 aware of the strength of its brand may show some strong resilience
and would prefer have these people as testers or specs validator than active
partners but that's an other story that repeat itself when an intitution is
gaining power (exception: IETF which is surprisingly democratic after all
these years, I am still amazed of that)

I am writing an article about CSS, XSL and DSSSL and more and more I dig in
the subject more and more these three things are very linked. Because the
same people are involved in these projects and also because these three
languages share the same goals and in a certain way could compete for our
mind share.

I think that the real issue here is not so much the piece of paper (a spec
is not more not less than that until it takes life with a real
implementation). It is more the fact that W3 is now a well known brand with
as much power as Microsoft in a certain part of the population. To say 100%
standard compliant is today a good marketing tool because the market wants
that. To say that it is 100% compliant with W3 is even more credible. The
whole point is here, W3 has power. Power to make or destroy. Power given by
that brand. As all institution they may use their power in one way or an
other. If W3 do not represent the interest of the user community of the
interest of small guys (sometime not having the money to pay for membership,
or having the money but not for all travel expenses and believe me, it can
cost you a lot just to be _there_ enough to influence, so you can make your
own deduction of who can influence the most). W3 do not necessarily
represent the end user community interests but first and foremost, its own
members interests (even if some W3 employees wants to and have good
intentions for these end users). So, W3 today has more power and if you look
closely to recent history, even Microsoft has to deal with now. W3 organism
will act for its own interest and that is the interest of its members. And
just look at this list historical track and you may notice as a good
historian that sometime the relationship is not necessarily two way (like
IETF) but more a like a group begging W3 to listen to their voice and no
official channel or mechanism to include non paying members (we should not
forget that it is not their business model) nor any voting forum open to
free participation (for all kinds of good reasons I am sure). So contrary to
DSSSL history where ISO only created very little action. XSL story started
with W3 created at least 5 to 10 implementations or work in progress from an
incomplete spec. This is incredible influence power (mostly based on the big
HTML base out there). If we wait for client side, you can guess who will
benefit (do you have doubts about it?) Waht will happen to these 5 to 10
early workers. Only help big guys reduce their expenses by having cheap
testers? Or have these people introduce their work on the market and maybe
in the log run have a chance to change the imbalance of power we have with
big companies. More than that, maybe have a chance to accelerate a bit the
introduction of this technology. In fact, Guy, these guys have more to loose
if XSL is not out here in the market before the end of the year than big
guys who don't give a dam anyway, they _have_ the market, so why rush? It
only entrepreurs or individual who have to gain somethin because they do not
_have_ the market but want to get a _piece_ of it. And these liliputians
often force big guys to move.

I do not think that people proposing a XSL spec split are in a go for an
other tea party :-) but try to make a living or believe in a cause or more
simply, have a job do to and has to deal with concrete stuff. It seems, that
using HTML as a rendering language is corresponding to actual market needs.
Having done the job with either DSSSL and XSL and I can provide to all real
examples like a DSSSL document and a XSL document doing the same job. Even
more, with improved version (DSSSL macros) could produce reasonable readable
scripts without too much ()))) stuff. I should say however that XSL
templates could be more intuitive for HTML developers especially the
"template" tag content and not necessarily the "Match" construct which can
be as confusing as a PERL construct, anyway not necessarily more simple than
a DSSSL construct. So here it is not so much technical stuff but commercial
realities, competition among standardization organism, competition among
languages and mind share, interest of a few, etc... the real life in fact. I
know Guy, this is not necessarily what I want too, but its life.
If I support Paul's initiative it is to have things going and not get stuck
in moving sands (big guys own the market so, why rush anyway?) and have this
whole product and enthousiam that W3 created about XSL reach the beach as
soon as possible and force as you want too, big guys to move and reach to
goal you also want as strongly as we do.

Yours truly
Didier PH Martin

 XSL-List info and archive:

Current Thread