Forwarded message 1
Ian Hickson wrote:
>On Fri, 5 Dec 2003, Daniel Brooks wrote:
>
>
>>I'd like to make one minor request. Could you make the yellow notes at
>>the side of the page (class="issue") narrower? The body has a left
>>padding of 8em, but the width of the issue box is 12em. You probably
>>need to compromise at 10 ems or so for each, because 8em is too thin (oh
>>how I love the dom inspector :)
>>
>>
>
>You have a minimum font size set, at a guess. That makes it your
>responsibility. :-) (It works fine here.)
>
>
>
That is interesting. I didn't expect it to change the width of the box,
but I suppose it's understandable. I expected it to use the original
font size as the amount used in the computation. It's like rounding, you
don't round off at every step, only after the last one. Of course, you
could just call it an inch or two, and call it a day.
Anyway, about the actual content, I do have a suggestion or two. I'm
going to have to go from memory, because my original email got eaten by
mozilla.
In section 7.4 I'd like to see a seperate error code for the file picker
to indicate when the specified file is not found. Lumping it in with
other error coniditions will only generate confusion on the part of the
user. You can't say the field needs to be filled out, especially since
they would have to type in a value themselves to get that result. The
others fit even less well. I really do like the user defined error
status though. That's a really great idea.
in 7.6, when you say "reset just the relevant control to its initial
value," do you mean that the UA uses just the initial value specified in
the value= attribute of the HTML, or is it supposed to also look in the
file specified by the control's form's data= attribute? If the control
is in a repitition block that's listed in the data file, it could use
the value specified for that block, and if it's in a repitition block
not listed (some new rows have been added since the page was loaded)
then it could use the default for the template block, etc.
Also, the parts about the onformchange event aren't entirely clear. If
the onformchange event on the first form control changes the value of a
form control, does that cause another set of onformchange events, or are
they batched up somehow? I'd guess the event dispatcher would have to
have a list of nodes to give the event to, and if the handler it called
changed something, it'd have to add all the nodes that came before it to
the end of the list. The ones that come after can just wait. You'd also
have to check to see if any given node wasn't already farther down on
the list. For example, the second node makes a change, so the first is
added to the end of the list. The fourth node then makes a change, so
you need to make sure the first is listed at some point down the list,
so that you don't add it twice etc. Could be a headache.
Section 4.4.1 completely lost me. As I understand it, you're saying that
the onformchange event could take a stripped down form of javascript
(ecmascript) simply because declaritive dependancies are easier to
track. Why bother? The javascript interpreter that pretty much has to be
built into any modern browser can already do that for the full js syntax
(though perhaps that data is never exposed to the outside, and it'd
probably take just as long as running the code in the first place to
determine the dependancies, etc)
The one large thing I'd do differently is the template/repeated block
stuff. With the system here, you have to completely specify the form
twice, first for the template and then for the first repeated block.
What I'd like to do is be able to just put repeat="repeated" on an
element and not put any content in it at all. The UA would take care of
filling out the DOM with a copy of the template's children. On the other
hand, maybe that's exactly what you had in mind and I just missed it. It
also might be nice to be able to for a repeate block to be a copy of a
template that isn't a sibling.
Anyway, I think it's a great idea, even though it does make Page Info a
little harder ;) Well, maybe we'll call it more interesting instead.
db48x