Re: Article on Testing Web Apps

> OK, a merciless savaging follows.

I guess I asked for it...

>  Don't take this the wrong way: I
> wouldn't take the time and effort to write this if I didn't think
> your article was worth it.

No, I don't take it the wrong way, and agree with your criticisms.  I
need to catch some sleep and deal with some other things before I can
update the article though.  Probably late tonight I'll have an improved
version up.

While I did edit it pretty carefully, it's the first draft anyone else
has seen.  I 
really do want it to be good so I definitely want feedback from experts.

I contemplated giving a specific example of a DTD, rendering into HTML
(with
the "<'s" and such escaped as entities) the DTD for a simple document
with a couple of elements and attributes, and explaining what it all
means.  I thought this might scare some people away though.  I know much
of the article is technical, and here I shy away from being specifically
technically, and that's not very consistent.

Wouldn't it be great if browsers summarily dropped invalid documents in
much the same way the IP protocol drops packets with bad checksums? 
Sure, it would have required more effort on everyone's part to get the
web started, but we'd all be better off now.  Imagine if the network
layer operated on a best guess as to what the data transmitter really
meant to say...

How should I argue that one should choose validation over prejudice
towards a
given browser?  

I have some advantage with the particular audience that this site
is aimed for; the Linux community uses lots of browsers that have
difficulty viewing
sites that are particular to some browsers - there's still a vocal Lynx
userbase!  A web application developer who uses Linux as both a server
and a desktop will be pretty sympathetic.

But really the site should be just as useful to Microsoft IIS developers
as Apache & Python developers.  I'm not sure what to say.  I do give the
section "Web Application Design Basics" which is meant to amount to a
boot camp on making sure your site isn't a terrible place to be.

> Check posts to the list over the past few days. This appears to be
>    a bug in the validator.

even so, these were all Linux sites I was validating.  I thought it was
pretty funny that it complained about a windows charset being missing. 
This suggests that these folks aren't the purists they claim to be,
maybe they had some web designer use a windows WYSIWYG tool to make
their site.  Or is that not the problem, but something just got screwed
up and the message is completely unrelated to what was in the pages?

I couldn't find ab and zb.  Do you have URLs for them?

> >    Tilting at Windmills for a Better Tomorrow.
> 
> I like it.  Succinct yet accurate!

I've used no other sig since the Spring of 1985.  It's a simple way of
expressing the
my belief that impossible should not stop me from doing what I feel is
right.

Like trying to convince everyone on the net to produce valid HTML
documents :-/ - or the overarching purpose of the Linux Quality site,
that Free Software should be of high quality too.  While there are
stellar examples of quality and fine architecture among GPLed programs,
there are some resounding thuds too.

I'm an avid reader of the Risks Forum.  If you're not familiar with it,
the
web version is at:

http://catless.ncl.ac.uk/Risks/

It's what turned me into such a zealot for software quality - and such
an uneasy
passenger when I fly on modern aircraft.

Regards,

Mike
-- 
Michael D. Crawford
GoingWare Inc. - Expert Software Development and Consulting
http://www.goingware.com
crawford@goingware.com

  Tilting at Windmills for a Better Tomorrow.

Received on Monday, 5 March 2001 06:59:50 UTC