RE: [xsl] Future of XSL Stylesheet Writing?

Subject: RE: [xsl] Future of XSL Stylesheet Writing?
From: "Scott Trenda" <Scott.Trenda@xxxxxxxx>
Date: Wed, 26 Sep 2007 19:50:35 -0500
(Note: semi-theoretical ramblings to follow. I may be wrong. :P)

I think I see what you're getting at with the TeX/CSS arguments, and
it's different here - with those and several other styling languages,
where you're taking a near-static input document and making it display
in a specific format. It's just a matter of time before someone makes
the process easier, and not long after that, someone figures out how to
make it dead simple. Sure, you _could_ write SVG paths by hand and from
memory, but why bother when you can just click and drag in
Inkscape/Illustrator? But remember, your data going into those
processors is, often by nature, static or rigid in its structure.


The problem with your proposed situation lies in the nature of
"WYSIWYG". In the case of XSL-FO, it may be possible, because you can
actually _see_ the effects of your output, and manipulate the
stylesheet's results visually. At least with a static input document,
you could.


The problem is, heavy-duty real-world XSLT stylesheets are usually set
up to handle a large range of input documents, and produce an output
document with a well-defined structure based on its input document's
structure and content. Since the structures of stylesheets created
through WYSIWYG tools are highly dependent on their sample input
documents, the range and accuracy of those stylesheets will only be a
matter of how many input documents the tool was provided to start with.
To produce robust XSLT, such a tool would have to define some sort of
pattern and handler syntaxes for the developer to generalize behaviors
for further input documents, and at that point, it would basically be
trying to recreate XPath and XSLT. Redundant, no?


Since the "styling" involved in XSLT is more of a logical
rearrangement-type styling than anything, the most useful tool for
creating XSLT is a solid view-input/edit-stylesheet/view-output
heuristic tool. The best editors on the market simply provide a fast,
convenient wrapper to this basic approach, and speed up the editing
process through auto-completion, XPath suggestion, step-by-step
debugging, and quicker access to real-world next-steps you're likely to
encounter.



Keep with it. Even if you don't end up using XSLT very often in your
day-to-day work, the functional-programming-oriented mindset will
eventually spill over into the other languages that you use frequently,
and you'll find yourself thinking outside the box by default there
because you got used to it in XSLT. (Either by choice or not. :P) And
when you encounter a new language, you'll be able to approach it from
the ground up, since XSLT (1.0) doesn't offer you much other than the
basic structures that can be used to build legions of others. It's worth
it, definitely.


~ Scott


-----Original Message-----
From: Steven Janoff [mailto:Steven.Janoff@xxxxxxx]
Sent: Wednesday, September 26, 2007 7:16 PM
To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
Subject: [xsl] Future of XSL Stylesheet Writing?

Hi,

Newbie here, well-trained recently in the XSL arts, with a
career-related question.

Been in technical publishing and related fields long enough to see
several generations of publishing solutions where "hot" skills go cold
or cool.  E.g., TeX coding skills replaced by secretary using Word;
HTML/CSS hand-coding skills replaced by graphic designer using
Dreamweaver.  And so on.

Now I'm knee-deep in XSLT/XSL-FO stylesheet writing.  I've wondered how
long these skills would be "hot" before being replaced by
much-less-skilled workers using a WYSIWYG XSL editor to create
stylesheets, without knowledge of the underlying XSL code.  And I see
the recent announcement of the first such tool (or the first I've heard
of), primarily applied to visual FO development.

How many years do you think it will be before the skills celebrated on
this list -- writing XSLT/XSL-FO stylesheets the "old-fashioned" way,
understanding the code -- will be supplanted by the scenario described
above, as happened with the earlier tools?

Will these skills serve me for a number of years (5? 10?), or will I be
looking for the next suite of tools to learn in just a few years as
grandmothers around the world start creating PDFs in XSL-FO at the push
of a button?  That's an exaggeration, but you know what I mean.

I've written this post in at least 6 ways (some much longer, some
shorter), but the fundamental question is the same:  Will these skills
be bankable in 5 or 10 years?

Thanks for your honest assessment here.  No one on the list can be
expected to predict the future, but the vast wealth of background among
you suggests that an "educated guess" from this list is about as close a
prognostication as you can get to what will actually happen.

Steve

--
Steve Janoff
Information Manager, Specialty Engineering
General Atomics Aeronautical Systems, Inc.
Tel. (858) 312-3255 (New number)
Fax (858) 312-4668 (New number)
Steven.Janoff@xxxxxxx

Current Thread