OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] Tim Bray on "Which Technologies Matter?"

[ Lists Home | Date Index | Thread Index ]

> > Why didn't anybody mention Adobe's Photoshop and Premier? They're both
> > famous for their complexity and impenetrable interface, and they both
> > dominate their respective niches. Photoshop, in particular, defeated some
> > very significant opposition. Only Paintshop Pro seems to be still hanging
> > in there ... and it sells to a completely different price/performance need.
>Oooh, let me think - in a lot of specialist markets the most complex tool
>wins since all the users are very results-focussed; they prefer to spend
>effort learning a complex interface to something powerful rather than a
>simple interface to something less powerful.
>Things used by programmers come to mind... Perl is very complex compared to
>Python, and people who use it tend to revel in the complexity and try to come
>up with more and more obscure tricks for getting things done.

Aha, finally something on xml-dev I'm qualified to comment on--

I find Photoshop brilliantly straightforward, and have, since I first 
tinkered with version 3.0, without instruction, as a 16-year-old.  This 
year I discovered Perl and XSLT, two languages that let me do useful 
things, fast, with minimal training.  Granted, there is power and 
complexity in all three that I will never use or care about.  But none of 
them (least of all Photoshop) has what I would call an "impenetrable 
interface."  What's critical here is to understand the difference between 
complexity and barriers to entry.  Perl may be extremely complex, but it 
has the lowest barriers to entry of any procedural programming language 
I've used since Commodore64 BASIC (granted, I haven't touched Python).  I 
looked at some code, read some online tutorials, and got to 
work.  Photoshop is the same story: I learned that interface by looking at 
the toolbar buttons and asking, "now, what is a tool that looks like this 
likely to do?"  So I learned the paintbrush, the airbrush, and more as I 
went along.  Sure, you can also do complex per-channel masking with 
path-based selections, but you don't need to know that to get things done.

I would never have touched computers if I had to memorize first and be 
productive later.  We got a Mac my junior year, I spent a day getting 
familiar with the desktop metaphor, and I was productive.  I hear you can 
do great things with vi, but I've never bothered to learn what they are 
because the interface to all the goodies is invisible.  Instead, I can use 
BBEdit or UltraEdit, both of which have incredible features *and* expose 
them to the user through a GUI.  Give me technology with a lot of power, a 
simple interface to what's simple, and a more advanced interface to what's 
more complex.  This is what Photoshop does, and I think it does it admirably.

This ties in with the thread, "XPath 2.0 - how much of XQuery should it 
include?".  As of last August, I knew some HTML and some Perl.  In 
September I was asked to create some auto-generated web reports--"and, oh, 
we'd like them in PDF format."  So it was time to learn XML, XSLT, and 
FOP.  By October, I knew something extremely useful, because the barriers 
to entry to XML, XSLT, and even to some extent XSL-FO, were low.  Just 
reading xml-dev scares me, because I get a sinking feeling in my stomach 
that the barriers to entry for the next round of all this stuff are going 
to be very, very high.  When a beginner, looking to do simple things, can't 
find a simple tutorial online clearly demonstrating the simple parts of a 
more complex spec, something is wrong.



News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS