From - Wed May 02 19:18:20 2001 X-Mozilla-Status: 0001 X-Mozilla-Status2: 00000000 Path: secnews.netscape.com!not-for-mail From: Bob Crawford Newsgroups: netscape.public.mozilla.general,netscape.public.mozilla.jseng Subject: Re: New javascript error dialog - BAD! Date: Sun, 05 Mar 2000 21:58:07 -0600 Organization: Another Netscape Collabra Server User Lines: 215 Message-ID: <38C32CCF.92AEAB49@io.com> References: <38C31C16.956685FE@io.com> <38C320E3.4EE546A2@netscape.com> NNTP-Posting-Host: cs9351-169.austin.rr.com Mime-Version: 1.0 Content-Type: multipart/alternative; boundary="------------1A5AC1602B430BA736D3DAA4" X-Mailer: Mozilla 4.7 [en] (WinNT; U) X-Accept-Language: en,pdf To: John Bandhauer Xref: secnews.netscape.com netscape.public.mozilla.general:16925 netscape.public.mozilla.jseng:1389 --------------1A5AC1602B430BA736D3DAA4 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit John Bandhauer wrote: > Bob Crawford wrote: > > > > I saw this for the first time in build 2000030416: > > 1- My site (slow server) dropped my .js file, so a javascript > > procedure was not available when needed. Things happen. > > > > 2- However, the following alert appeared: > > > > "This error may indicate that the site has > > not been updated for Mozilla. Please > > contact the site administrator, or see > > http://developer.netscape.com/mozilla" > > > > This is the wrong thing to say, and positions Mozilla as > > a "Problem Child." > > This will almost certainly result in misunderstandings > > and confusion, and will not cause developers to rush to support > > Mozilla. > > The alert should say that there was an error loading the > > javascript for the page, rather than attacking the site's > > designers. > > (Note - if a site does not take Moz into account, it is MOST > > likely to manifest in bad browser-sniffing, which will not be > > caught by said trap.) > > I disagree. Failure to load a .js file is an unusual edge case. I would not say that. The net is a tempermental beastie and it is not at all unusual for a file to temporarily go -missing-. I would not refer to this as an "unusual edge case." The net is supposed to be fault tolerant - you must PLAN a safe failsafe if a needed file fails. A missing '.css' just renders like crap without the need to pop an alert box. It might actually be cool if an alert popped up stating "SOME STYLE DEFINITIONS WERE UNABLE TO BE LOAD, THIS DOCUMENT MAY NOT DISPLAY CORRECTLY" or some such... > The more likely causes of JS errors are from JS code that won't > compile or run in mozilla. And the most likely cause of that is > (as you suggest) bad sniffing. I Disagree - the MOST likely cause is simply buggy code, or missing files (or crreped-out files due to dropped packets...).. there IS a lot of it about... What I meant to suggest above is that the most common manner in which a site will fail to accomodate Moz is this: Bad browser detection. Much bad browser detection will make it through the interpreter just fine (Version string checking, which is the most common method of browser detection used.) This may NOT cause the dreaded calls to .all or .layers, but simply result in a dumbed-down page or a "this browser not supported page." I would might welcome an alert which triggered specifically on calls to .all or .layers, but to trigger it for every Javascript error (or every nonexistant function) would be uncalled for, misleading, and annoying.. Also, it might be better to point out that the site does not support the W3C spec, rather than the particular browser... put the 'blame' clearly on non-standard-compliant web-coders, rather than some pain-in-the-butt browser which must be specifically accomodated (this is what this next-level stuff is supposed to obviate...). > I'd bet most problems will be in > pages that get served with 'document.all' or with references to > 'layer'. These are exactly the cases where we'd like people to > know that the maintainers of the site they are looking at should > be urged to update for compatibility with mozilla. See above - what should be emphasized is that the sites need to comply to the W3C STANDARDS, not an upstart-browser. The perception that Moz is a problem which must be specifically coded for will put a stake through this thing's heart faster than you can say "ouch!" > Otherwise > people are most likely to blame mozilla for not properly > rendering the site. We are trying to preempt a lot of misdirected > bug reports and avoid the general perception that mozilla is at > fault. I certainly agree on many, many points.. see http://www.w3cdom.com/detect.html > > (By The Way... what ever happened to the javascript console?? > > We need this...) > It is on its way. It is not going to make it into beta1. Sigh.. I'll try to be patient! > John. -------- Bob Crawford - crawdad@io.com Disclaimer: The above comment is in no way intended to be mean or nasty. Sometimes enthusiasm for one's viewpoint is confused with rancor. This is not intended, please do not flame. --------------1A5AC1602B430BA736D3DAA4 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit John Bandhauer wrote:
Bob Crawford wrote:
>
> I saw this for the first time in build 2000030416:
> 1- My site (slow server) dropped my .js file, so a javascript
> procedure was not available when needed. Things happen.
>
> 2- However, the following alert appeared:
>
>           "This error may indicate that the site has
>           not been updated for Mozilla. Please
>           contact the site administrator, or see
>           http://developer.netscape.com/mozilla"
>
>       This is the wrong thing to say, and positions Mozilla as
> a "Problem Child."
>       This will almost certainly result in misunderstandings
> and confusion, and will not cause developers to rush to support
> Mozilla.
>       The alert should say that there was an error loading the
> javascript for the page, rather than attacking the site's
> designers.
> (Note - if a site does not take Moz into account, it is MOST
> likely to manifest in bad browser-sniffing, which will not be
> caught by said trap.)

I disagree. Failure to load a .js file is an unusual edge case.

     I would not say that. The net is a tempermental beastie and it is not at all unusual for a file to temporarily go -missing-. I would not refer to this as an "unusual edge case."
     The net is supposed to be fault tolerant - you must PLAN a safe failsafe if a needed file fails.
     A missing '.css' just renders like crap without the need to pop an alert box. It might actually be cool if an alert popped up stating "SOME STYLE DEFINITIONS WERE UNABLE TO BE LOAD, THIS DOCUMENT MAY NOT DISPLAY CORRECTLY" or some such...
 
 
The more likely causes of JS errors are from JS code that won't
compile or run in mozilla. And the most likely cause of that is
(as you suggest) bad sniffing.
     I Disagree - the MOST likely cause is simply buggy code, or missing files (or crreped-out files due to dropped packets...).. there IS a lot of it about...
    What I meant to suggest above is that the most common manner in which a site will fail to accomodate Moz is this: Bad browser detection.
    Much bad browser detection will make it through the interpreter just fine (Version string checking, which is the most common method of browser detection used.) This may NOT cause the dreaded calls to .all or .layers, but simply result in a dumbed-down page or a "this browser not supported page."

    I would might welcome an alert which triggered specifically on calls to .all or .layers, but to trigger it for every Javascript error (or every nonexistant function) would be uncalled for, misleading, and annoying..

    Also, it might be better to point out that the site does not support the W3C spec, rather than the particular browser... put the 'blame' clearly on non-standard-compliant web-coders, rather than some pain-in-the-butt browser which must be specifically accomodated (this is what this next-level stuff is supposed to obviate...).
 

I'd bet most problems will be in
pages that get served with 'document.all' or with references to
'layer'. These are exactly the cases where we'd like people to
know that the maintainers of the site they are looking at should
be urged to update for compatibility with mozilla.
See above - what should be emphasized is that the sites need to comply to the W3C STANDARDS, not an upstart-browser. The perception that Moz is a problem which must be specifically coded for will put a stake through this thing's heart faster than you can say "ouch!"
Otherwise
people are most likely to blame mozilla for not properly
rendering the site. We are trying to preempt a lot of misdirected
bug reports and avoid the general perception that mozilla is at
fault.
I certainly agree on many, many points.. see http://www.w3cdom.com/detect.html
> (By The Way... what ever happened to the javascript console??
> We need this...)
It is on its way. It is not going to make it into beta1.
Sigh.. I'll try to be patient! 
John.
--------
Bob Crawford - crawdad@io.com
Disclaimer: The above comment is in no way intended to be mean or nasty. Sometimes enthusiasm for one's viewpoint is confused with rancor. This is not intended, please do not flame.
 
  --------------1A5AC1602B430BA736D3DAA4--