[thelist] What is wrong with this site?

Jeff Howden jeff at jeffhowden.com
Tue Aug 19 14:23:50 CDT 2003


tim,

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> From: Tim Fountain
>
> > Jeff Howden wrote:
> >
> > wrong. a web browser is a tool for navigating the web
> > and viewing documents available online, offline, etc.
>
> The problem is, it's both. It's a tool for navigation
> but it's also the tool developers use to test pages.
> Therefore by protecting users from mistakes you're also
> hiding them from developers, which in turn is bad for
> users.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

you're assuming the browser was built originally as both a viewing tool and
a debugging tool.  that assumption is incorrect.  it's first a viewer and by
virtue of being the only suitable tool for the job, also the debugger.

i'll agree that hiding mistakes from the developer is bad.  but i'll
disagree that hiding them from the user is bad.  on some things there are
toggles to make it easier for developers to move around, isolate and
identify problems, etc.  you won't find all the necessary tools available
within the browser though as debugging/development is not its primary
function.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> As an example, I was browsing a high profile e-commerce
> site a month or two ago, I think it was lastminute.com.
> Their server was sending pages back as text/plain, and
> in my browser of choice this made their site completely
> unusable. I don't know how long the site had been like
> this (it's fixed now), but it could have been a while.
> If all browsers displayed the content sent as
> text/plain as just that I'd care to bet the problem
> would have never been there, but if it was it would
> have been noticed very quickly.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

agreed, but it could also be argued that the 95% of the audience to the site
would not have been able to find what they were looking for because of
something really simple being set incorrectly.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> The reason we have standards in the first place is to
> try and tie all browsers together so they behave in
> roughly the same way.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

*yawn*  standards is not what this conversation is about.  please don't
cloud it with that.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> However if all browsers were 100% standards compliant,
> as long as your page was standards compliant as well you
> could be pretty sure it would work the same in all
> browsers. It's impossible for developers to check sites
> on _all_ browsers, there are hundreds of them.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

*yawn* more standards nonsense.

you contradict yourself in the passage above.  furthermore, the reality is
that not all browsers are 100% standards compliant.  in fact, there isn't a
single browser on the market that's 100% standards compliant; not one.  even
the ones that are close have added their own proprietary garbage to html,
css, etc. making the reality of 100% standards compliance getting the job
done a serious joke.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> > if, as a developer, i make an innocent mistake, the
> > user's shouldn't have to suffer. the browser should
> > be smart enough to work around the mistake (providing
> > it's not too big).
>
> But how? Fixing text/plain served content isn't a good
> example here, so say you've got a messy HTML page with
> lots of missing </table> and other tags to the point
> where it's not obvious how it's meant to look. The
> browser may "guess", but one browser's guess isn't
> necessarily going to be the same as another browser's,
> which lead to a readable page in one and an unreadable
> page in another. As I said earlier, because you can't
> test in _every_ browser, this is A Bad Thing.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

validating your code would solve that problem immediately.

a browser displaying a blank page because a closing table tag is missing is
exceedingly unhelpful to a user.  let the browser fail gracefully for the
developer that's either html-illiterate or too lazy to run the code through
a validator, but don't punish the user.

.jeff

------------------------------------------------------
Jeff Howden - Web Application Specialist
Resume - http://jeffhowden.com/about/resume/
Code Library - http://evolt.jeffhowden.com/jeff/code/




More information about the thelist mailing list