New Dojo Site--Most incompetent ever?

[David Mark]

David Mark wrote:
SteveYoungGoogle wrote:
David Mark wrote:
SteveYoungGoogle wrote:
S.T. wrote:
On 3/8/2010 1:04 AM, David Mark wrote:

[...]

The problem is the code itself. The code is large. There is disturbingly
faulty logic in the core of dojo itself (some of it discussed in this NG
archives).

It's the tip of the iceberg. Wait until you see my review of Dojo 1.4
(coming soon to cinsoft.net). Of course, it looks very much like Dojo
1.3 as nobody over there touches the core, which is understandable once
you grasp the depth of their misunderstandings (who wants to fiddle with
a foundation that they can't even explain).

The core has many dependencies and so trying to adjust that, with so
many things that require it, is a risk.

I assume by dependencies, you mean the modules built on top of the core.
Yes, there are a lot of them (they tend to throw anything and
everything in there, regardless of whether it actually works). But if
you know what you are doing (and ostensibly _somebody_ involved with
that project does), it's not that bad. Lots of interdependencies in the
core itself, but that's more of an inconvenience than a real issue. The
thing is, most of the individual modules are pretty basic. Their
biggest claim to "fame" is that they have a bunch of widgets, supposedly
by IBM, but in reality the work of someone called "bill". They are some
pretty awful widgets and somehow their focus is adding more "themes"
(they have all of two, with a distinct style sheet per widget), rather
than making the widget code competent.

Anyway, I did most of the heavy lifting already (particularly in the
core). But then it became apparent that the other contributors didn't
see the whole as a mess and only wanted to discuss one tiny patch at a
time (which would mean completing the transformation would take
approximately fifty years).

OTOH, the problems won't go away
by ignoring them.

Ain't that the truth.

Fixing the problems sounds like a good idea, but then
there would have to be somebody capable of doing that, and if it is
going to be done by the original authors, then much learning should take
place prior to doing that (or different mistakes will be made).

Exactly! But, oddly enough, the original authors don't see it that way.
In fact, the very idea that so many changes (in code and
understanding) could be needed caused a complete nullification.

Regaring your review, I would like to suggest the following document:

http://www.jibbering.com/faq/notes/review/

Yes, I've read it. IIRC, I agreed with much of it.

I'd also like comments on how it can be improved.

I'll see if I can make some time for that, but I am swamped at the moment.

The Dojo site is a fine incompetence exemplar, but so are others.

I was referring to the Foundation site. I've never seen anything quite
so resolution-challenged. It's like the developer(s) had their browsers
maximized at a relatively high resolution during the entire testing
process (assuming there was a testing process). And the fact that it is
throwing exceptions in brand new browsers (e.g. Opera 10) goes to show
how anything based on Dojo (even basic pages by the authors/supporters
of the library) falls apart in short order (requiring a huge upgrade
ordeal). There's a business opportunity in there somewhere as I know a
lot of companies are currently stuck in such endless cycles behind their
firewalls. I'm working on that as part of my new site. I can offer a
far more future-proof Dojo or fix the one they have.

A bad corporate website is not nearly as bad as a large corporate
software project failing. Half a million dollars is not something that
should be thrown in the trash.

That's what I'm talking about.

 

[David Mark]

Good example, it is also important consider characters that need to be
percent-encoded, such as: (, ), %.

And that's just for A href.

If the HTML is valid, the program can be focused more on the
requirements problems and not how browsers handle errors.

And isn't that the pitch of these things? They want to save everyone
from the rigors of front-end development, yet their developers are
almost exclusively from the "just gettin' stuff done" school where
anything goes, so long as they can see it "work" in their test browsers.
Of course, the marketers and shills put a different spin on it. They
are just being "pragmatic" (and some actually _believe_ that BS). I
guess it's not a lie if you believe it.

 

[Garrett Smith]

[...]

I was referring to the Foundation site.

Yes I know and I have seen the resizing issue there. There are many
sites that are worse, breaking functionality. Pubmed, for example. I
love the content but what an awful UI.

The web is a mess.

I've never seen anything quite

so resolution-challenged. It's like the developer(s) had their browsers
maximized at a relatively high resolution during the entire testing
process (assuming there was a testing process). And the fact that it is
throwing exceptions in brand new browsers (e.g. Opera 10)

Firefox 3.5 - "An invalid or illegal string was specified" code: "12"

I get that every time I click on a tab header.

And in Safari 4 - "Error: SYNTAX_ERR: DOM Exception 12"

I don't have the energy to dig through their obfuscated code to find
exactly what it is, but it appears to be dom-related.

The iframe to try to fix the history has some odd markup:
http://www.dojofoundation.org/media/documents/iframe_history.html
<meta http-equiv="Content-Type" content="text/html; charset=utf-8"></meta>

 

[David Mark]

David Mark wrote:
Garrett Smith wrote:
David Mark wrote:
SteveYoungGoogle wrote:
David Mark wrote:
SteveYoungGoogle wrote:
S.T. wrote:
On 3/8/2010 1:04 AM, David Mark wrote:

[...]

I was referring to the Foundation site.

Yes I know and I have seen the resizing issue there. There are many
sites that are worse, breaking functionality. Pubmed, for example. I
love the content but what an awful UI.

That one was pretty much the pits in my book. I've seen some bad sites,
but when you've got a single page that reaches for the stars with
Ajax-ified sliding of content bits and fails to even allow reading the
content, you've got a candidate for worst site in history. Even a basic
page with no style, no script, etc. would beat that. Yet, somebody
somewhere signed off on that thing as "cool" and would likely have
sniffed at doing a basic site with normal navigation. I think most Web
developers try to please themselves (and must be extraordinarily easy to
please).

The web is a mess.

That's an understatement. And you should see the train wrecks behind
most of it.

I've never seen anything quite

Firefox 3.5 - "An invalid or illegal string was specified" code: "12"

Yes and that browser was certainly in widespread use (and was one they
asserted to "care" about) at the time that page was created. Makes you
wonder what sort of horror show you would get in a browser they don't
care about. It simultaneously gives Open Source and the Web black eyes.

I get that every time I click on a tab header.

Yes, you wonder if anybody bothered to test the page at all. Wonder how
it will fare in FF4. I haven't even tried it in IE (lately), but I
imagine it can be coaxed into throwing errors there as well.

And in Safari 4 - "Error: SYNTAX_ERR: DOM Exception 12"

Now I know that's one browser they "care" about. In fact, it's the only
Safari they claim to "support". Makes you wonder how anybody gets
sucked into using Dojo to create applications when it is clear the
people promoting it are out of their league creating a simple document.

I don't have the energy to dig through their obfuscated code to find
exactly what it is, but it appears to be dom-related.

Yeah, they apparently don't have the energy either. You know I reported
these problems a long time ago.

The iframe to try to fix the history has some odd markup:
http://www.dojofoundation.org/media/documents/iframe_history.html
<meta http-equiv="Content-Type" content="text/html; charset=utf-8"></meta>

LOL. The whole idea of using an IFrame to "fix history" so that
visitors won't be "inconvenienced" by navigation is ludicrous beyond
belief, but such thinking is blasphemy to this bunch. Then I guess when
you have 500K pages with 50 http requests, navigation can be a bit
sluggish. They've designed themselves into a box and can't punch their
way out of it to save their lives (or reputation). IMO, people who
think like that belong in rubber rooms, not cubicles (and they sure as
hell shouldn't be let near company Websites).

 

[Richard Cornford]

I worry about what the marketplace has specified, not a W3C
decade-long adventure in producing a "Recommendation" that
sometimes is, sometimes is not followed.

An unlikely example. I'd agree it's best to avoid Hx tags inside
spans, but objected to a scathing condemnation of Dojo's site
because they had a block inside an inline and had the audacity
to allow CSS to ensure the user sees the intended effect.
Suggesting they swap the absolute positioned span to an absolute
positioned div is fine. Mocking them because they haven't bothered
to was absurd.

<snip>

It is not as absurd as it may seem. It does directly reflect on Dojo
as a scripting project, and on its author's, experience/understanding
in relation to DOM scripting.

It has long been known that when presented with structurally invalid
mark-up the error correction applied by web browsers can result in
(sometimes radically) divergent DOM structures (including structures
where some elements become the decedents/children of multiple elements
(which is not possible in a true tree-like structure such as the DOM
is supposed to be).

This is very significant for anyone who is trying to create non-
trivial DOM scripts (which is presumably something Dojo is attempting
to qualify as), and so should be understood by any individual involved
in such a task. The result of that understanding is usually an
appreciation of the value of using structurally valid mark-up (as
represented by the participants in this group[1]), and so observing
the use of invalid mark-up implies a failure to understand this point
(a shortfall of experience or knowledge).

Here is the URL of post of mine from 2005 that included a number of
test cases (based on real-world examples from questions asked on this
group):-

<URL: http://groups.google.com/group/comp.lang.javascript/msg/2672da2374ec16c2
Some of the handling of the test-case code has changed in more modern
browsers (showing that error-correction in any given browser is not
even stable over time), but it is still the case that you get to see 3
different DOM structures in, say, Firefox, IE and Opera.

Richard.

[1] There may be obsessive, anal, etc. individuals who insist on the
use of valid mark-up because following rules (regardless of any need)
makes them feel better, but there are practical, demonstrable and
significant consequences of the use of structurally invalid mark-up in
relation to DOM scripting. So the motivation for insisting on the use
of valid mark-up elsewhere (in other groups, in relation to other
fields) may not be entirely valid, but in the context of DOM
scripting, at least, the arguments for structural validity in mark-up
are substantial. Here, only the ignorant and/or inexperienced (or in
VK's case, barking mad) would be happy to see structurally invalid
mark-up.

 

[S.T.]

That is *one* (1) example to support your thesis. Browsers with any
market share commonly support at least 96 counter examples.

<http://bednarz.nl/tmp/entref/>

Point taken.

Perhaps naively, I thought the format had to be &xxx; (with the trailing
semi-colon) and figured the odds of me naming a GET something like
'divide;' was nil. I guess there is somewhat of a risk of naming
something 'pound' or 'para' though.

 

[David Mark]

<snip>

It is not as absurd as it may seem. It does directly reflect on Dojo
as a scripting project, and on its author's, experience/understanding
in relation to DOM scripting.

Which is virtually nil (still). I was just reading some reviews of
their 2007 model. Nothing has changed. If anything, the delusions have
become more dramatic.

[...]

This is very significant for anyone who is trying to create non-
trivial DOM scripts (which is presumably something Dojo is attempting
to qualify as), and so should be understood by any individual involved
in such a task.

Yes, when the Dojo authors can't create a very basic Website (they've
had myriad shots at it over the years) with their own wonder-scripts,
anyone with sense must conclude they are barking up the wrong tree (i.e.
they should find a new hobby).

The result of that understanding is usually an
appreciation of the value of using structurally valid mark-up (as
represented by the participants in this group[1]), and so observing
the use of invalid mark-up implies a failure to understand this point
(a shortfall of experience or knowledge).

As evidenced, they fail to understand even the most basic concepts of
sound Web development. Furthermore, they can't seem to grasp
abstractions, preferring to demand empirical "proofs" for what should be
self-evident propositions. How they figure they can keep rewriting the
stupid thing in perpetuity, constantly changing the code to "keep up"
with the latest browsers, piling bad code on top of worse and discarding
older browsers (and visitors) on whims is beyond my comprehension. It
seems like they perversely (almost gleefully) strive to do everything
backwards and are quite unwilling to consider contrary points of view.
Small wonder the project has gone nowhere, but they seem to cling to the
idea that next year could be the one where everyone comes around to see
things their way.

 

[Dr J R Stockton]

In comp.lang.javascript message <[email protected]

Validating is a debugging tool

Agreed. It finds three types of bugs :
A. Those which you could find by testing & reading the page
B. Those which you would not find by testing & reading the page
C. Those which no-one will ever be affected by.

It is useful for A, very useful for B, and unimportant for C.

- that's it. It's not important if a page "passes" or not.

No doubt there are lengthy arguments about how critical validating is
to the future of humanity, but the real world uses validation for it's
useful purposes and stops there. ALT'ing every single IMG whether
useful or not is a fool's errand. Escaping every ampersand in a URL is
wasted time.

After realising that validation is useful for detecting errors, the step
which you should take next is to realise that it is much easier to check
a validator's output for important errors if the output is not cluttered
up by a large number of easily-fixed errors that you are not otherwise
bothered by.

 

[Eric Bednarz]

[…] but in the context of DOM
scripting, at least, the arguments for structural validity in mark-up
are substantial.

Now everybody who nevertheless uses a Firefox debugging add-on that not
only modifies the DOM, but does so in a way that breaks this structural
validity, raise your hand.

 

[David Mark]

[…] but in the context of DOM
scripting, at least, the arguments for structural validity in mark-up
are substantial.

Now everybody who nevertheless uses a Firefox debugging add-on that not
only modifies the DOM, but does so in a way that breaks this structural
validity, raise your hand.

I swore off that piece of crap long ago for that and numerous other
reasons. Never again. That stupid IETester does something similar,
which drove me nuts a little while back because it was causing a query
test to fail.