Flutterby™! : Supporting Mac IE

Next unread comment / Catchup all unread comments User Account Info | Logout | XML/Pilot/etc versions | Long version (with comments) | Weblog archives | Site Map | | Browse Topics

Supporting Mac IE

2001-08-20 18:24:36+00 by Dan Lyke 20 comments

Dori complained about the Umbra notice that they won't support Mac IE, which seems contradictory to her support of the Web Standards Project browser upgrade campaign. I still mean to try to figure out just whatinthehell Mac IE is sending in form text fields, Mars and others have complained that something it's doing ends up causing my formatting engine to double-space their comments, but I've no idea what that is.

[ related topics: Web development Web Standards Project - WaSP ]

comments in ascending chronological order (reverse):

#Comment made: 2002-02-21 05:32:32+00 by: jjg

This is primarily a test post to see if I get the same double-spacing thing, and also to see

what goes out in the POST request.

This is a second paragraph, which serves no purpose except to privde more text for the

test case.

Thank you for reading.

#Comment made: 2002-02-21 05:32:32+00 by: jjg

Okay, Mac IE is sending CR+LF at every line wrap. Your formatting engine seems to be interpreting them as paragraph breaks, rather than as simple line breaks. I see two ways to address this: either change your textarea tag to wrap="soft", which will still wrap in the input interface but not insert CR+LF in the data; or modify the formatting engine to interpret CR+LF as a line break. Of course, either of these may break functionality in other browsers -- what do I look like, a QA team?!?

#Comment made: 2002-02-21 05:32:32+00 by: jjg

Now I'm really confused. That time it worked fine!

#Comment made: 2002-02-21 05:32:32+00 by: jjg

One last thing: there should really be a link back to the home page on this one.

#Comment made: 2002-02-21 05:32:33+00 by: TheSHAD0W

That's because, in your 2nd posting, you never hit enter. As a result, you didn't get that CR+LF in there.

#Comment made: 2002-02-21 05:32:33+00 by: Dan Lyke

Aha! I thiok the issue is really with the Perl CGI module, which echoes back the CR/LF as a LF/LF, but the second time, when you had the cookie and the post wasn't getting stuffed back into a hidden field, it was getting sent as CR/LF.

I wonder why IE for Windows, which I've observed sending the CR/LF (and tried to account for), doesn't seem to have this problem.

#Comment made: 2002-02-21 05:32:33+00 by: Larry Burton

>> Dori complained about the Umbra notice that they won't support Mac IE, which seems contradictory to her support of the Web Standards Project browser upgrade campaign. <<

The difference is that the Web Standards Project browser upgrade campaign is about making the site at least usable to all browsers but just informing the users of non-compliant browsers that they are missing part of the formatting and directing them to where they can upgrade to a current standards compliant browser. The Umbra site just says, "Go away until you start using a browser we approve of." At least that's the way I see it.

#Comment made: 2002-02-21 05:32:33+00 by: Dan Lyke

Dori's Wedlog doesn't offer the option to Netscape 4.7x browsers which have JavaScript turned on (I'm willing to chalk up the SIGSEGV when CSS is turned on to a browser bug). I'd say that's the same thing.

However, I do realize that not everyone who links to the Web Standards Project upgrade page is taking such a "use our favorite browsers or go away" stance.

#Comment made: 2002-02-21 05:32:34+00 by: Dori

Dori's Wedlog doesn't offer the option to Netscape 4.7x browsers which have JavaScript turned on (I'm willing to chalk up the SIGSEGV when CSS is turned on to a browser bug). I'd say that's the same thing.

The best description of how the WASP feels about this issue is at Expectations and Misunderstandings.

In this case, Wedlog was written for one user: me. Anyone else wandering into it wasn't part of the target market, so I wasn't going to spend a lot of time working around bugs in their browser.

As to being allowed on the page with CSS turned off but JavaScript turned on, unfortunately, there's no way for me to test for that. And if I let people on the page with both JavaScript and CSS on, Netscape 4.x dies a horrible death. I figure it's better to be sent to a page with upgrade info than to crash.

This page, otoh, is for the general public. I still wanted to use XHTML and CSS without worrying about Netscape 4.x, so I coded it that way.

Please don't mistake my personal notes to myself for info about how the Web Standards Project says the web ought to be. I ain't got anywhere near that kind of authority in the group <g>.

But the message from Umbra is just plain rude and ignorant.

Dori
Backup Brain

#Comment made: 2002-02-21 05:32:34+00 by: Dan Lyke

Wedlog crashed the version of Netscape 4.x I was trying it with (it was an earlier one, whatever's the default on the boxes at work) without redirect when CSS was turned on. Thus the only effect of the redirect was to make the page unusable even when the browser could read it.

(Unless JavaScript was turned completely off, which I heartily endorse, but since you're one of the leading lights on JavaScript I'd hardly expect that you'd adopt my stance on that matter...)

As you say, that's a personal site and you can do whatever the heck you want with it, but to my mind any redirection based on browser detection is the same sort of snobbery that's shown in that Umbra message.

The explanations page attack on Dreamweaver isn't nearly specific enough, the spec for HTML3 (not 3.2) doesn't require a doc type. Yes, I have spent way too much time following the W3C validator implementor's rationale for testing various features.

#Comment made: 2002-02-21 05:32:34+00 by: Dori

Wedlog crashed the version of Netscape 4.x I was trying it with (it was an earlier one, whatever's the default on the boxes at work) without redirect when CSS was turned on. Thus the only effect of the redirect was to make the page unusable even when the browser could read it.

I think that you went there when I first started that blog. It should never crash anyone any more (hopefully!). If you're visiting with a browser that might crash, it should send you to the Web Standards Upgrade Page instead.

(Unless JavaScript was turned completely off, which I heartily endorse, but since you're one of the leading lights on JavaScript I'd hardly expect that you'd adopt my stance on that matter...)

If you absolutely have to use Netscape 4.x, I think that you should have JavaScript and CSS both turned off. You have a deeply broken browser, and at least with those "features" turned off you'll crash less often. Just my opinion.

But as that upgrade page says, there are lots of good browsers that aren't Netscape 4.x, from all kinds of vendors for all kinds of platforms. Any most of them don't have the bloat of Netscape 4.x, 'cause they don't throw in a mail reader, news reader, HTML editor, and kitchen sink along with simply reading web pages.

As you say, that's a personal site and you can do whatever the heck you want with it, but to my mind any redirection based on browser detection is the same sort of snobbery that's shown in that Umbra message.

I think that target audience is what's important. A commercial web site is completely different from a personal web site, in that an audience of many requires a completely different attitude that that of an audience of one.

For example, I still find Flutterby to be only barely useable. I dislike your choice of link colors (links are 330088 and vlinks are 000088, which look the same to me); it makes your site difficult for me to navigate. But , imo, it's a personal site, and it's yours to do with whatever you want.

If you were trying to sell me something, though, it's unlikely I'd buy it--your site's too confusing for me.

Umbra, otoh, is trying to sell me something (I think--I still have no idea what they do). And because of that, they've lost my business.

The explanations page attack on Dreamweaver isn't nearly specific enough, the spec for HTML3 (not 3.2) doesn't require a doc type. Yes, I have spent way too much time following the W3C validator implementor's rationale for testing various features.

Just out of curiousity: why would you want to validate against 3.0 instead of 3.2?

#Comment made: 2002-02-21 05:32:34+00 by: Dan Lyke

My experiences with the Wedlog crashes with Netscape 4.72 were on Wednesday. The browser crashed with CSS turned on, was redirected with JavaScript on and CSS off. So the only thing the redirect did was make the site unreadable in a situation where it would have otherwise been readable.

I still don't agree that there are viable alternatives to Netscape on the Macintosh. None of my non-computer literate friends who own Macintoshes will try IE again after reinstalling their OS to get rid of the crashes introduced with the install of IE. Opera is in beta, but not there yet. The solution to that problem will probably include most of them migrating off the Mac.

Obviously I tend to use Lynx and Opera[Wiki] a lot, both "standards compliant", but I still find myself often using machines I don't control with older Netscape versions.

Flutterby color usability: That's why the default is your browser defaults. The color schemes primarily exist because various "web designers" tried to impose defaults in broken ways and I wanted to read my site and still use the web, which meant that my site and not my browser defaults had to change. Of course now that Opera[Wiki] gives me real choice in that matter I care less, and the color schemes will probably eventually go away.

Why would I validate 3.0 versus 3.2? The Expectations and Misunderstandings says "For instance, a site that uses FONT tags and an HTML 3.2 DOCTYPE can be a perfectly valid, standards-compliant site." I haven't diddled with Dreamweaver[Wiki], but is it putting out 3.0? (I can't find the 3.0 spec on the W3 site, my note that it doesn't need a doc type is based on discussion on the W3 Validator mailing list about why the W3 Validator[Wiki] doesn't require a doc type).

#Comment made: 2001-08-24 21:38:37+00 by: Dori [edit history]

My experiences with the Wedlog crashes with Netscape 4.72 were on Wednesday. The browser crashed with CSS turned on, was redirected with JavaScript on and CSS off. So the only thing the redirect did was make the site unreadable in a situation where it would have otherwise been readable.

Hmmm... obviously, I haven't done that much testing with this situation, but I thought that if JavaScript was turned off in Netscape 4.x (as we discussed before), that meant that CSS was off automatically.

So, I thought the choices were:
- JavaScript on and CSS on: redirect.
- JavaScript on and CSS off: redirect.
- JavaScript off and CSS irrelevant: ugly and unreadable, but no crash.

With Netscape 4.x, the choices are between crashing or redirecting. I think that redirecting is more polite (considering there's nothing there that's relevant or important, anyway). From a user standpoint, I think that redirection gives them (possibly useful) info, while crashing does not.

I still don't agree that there are viable alternatives to Netscape on the Macintosh. None of my non-computer literate friends who own Macintoshes will try IE again after reinstalling their OS to get rid of the crashes introduced with the install of IE. Opera is in beta, but not there yet. The solution to that problem will probably include most of them migrating off the Mac.

Have they tried iCab?

And seriously: I have been using a Mac for many, many years. I have never, not once, ever, needed to reinstall an OS.

Telling Mac users to reinstall the Mac OS is similar to telling Mac users they need to edit their registry. It's stuff that works on Windows, but is never or rarely needed on the Mac.

For the folks with Macs who really want/need to use Netscape, personally, I recommend Netscape 3. Netscape 4 didn't add much besides half-hearted buggy attempts to do things; it's simpler to just do without them altogether.

Flutterby color usability: That's why the default is your browser defaults. The color schemes primarily exist because various "web designers" tried to impose defaults in broken ways and I wanted to read my site and still use the web, which meant that my site and not my browser defaults had to change. Of course now that Opera gives me real choice in that matter I care less, and the color schemes will probably eventually go away.

Well, I didn't touch anything on my side and the colors are different today (vlinks are "#551a8b" and links are "#0000ff"). So, I'll assume that something changed on your end and just say "Thanks!"

Why would I validate 3.0 versus 3.2? The Expectations and Misunderstandings says "For instance, a site that uses FONT tags and an HTML 3.2 DOCTYPE can be a perfectly valid, standards-compliant site." I haven't diddled with Dreamweaver, but is it putting out 3.0? (I can't find the 3.0 spec on the W3 site, my note that it doesn't need a doc type is based on discussion on the W3 Validator mailing list about why the W3 Validator doesn't require a doc type).

I don't know whether DW puts out 3.0. Regarding DW, I don't think that it (currently) cares about compliance to any spec at all. Regardng 3.0, I believe that 3.2 was the first HTML recommendation that the W3C ever produced, so I use that as a baseline. I haven't been able to find any published version of 3.0, either.

#Comment made: 2002-02-21 05:32:34+00 by: Dan Lyke

Whoops: Clarification:

  1. Netscape 4.x, JavaScript turned off, no matter what the CSS state, no problem.
  2. Netscape 4.x, JavaScript turned on, CSS off, redirect. Presumably the page should be visible in this state.
  3. Netscape 4.72, JavaScript on, CSS on, crash. (4.76 might have executed the redirect, I don't remember). Thus the redirect only happens in a situation where the page would probably have displayed anyway.

Re: Mac OS/Browsers: In my career, never having personally owned a Mac, I've spent more time attempting to fix the MacOS than any other operating system. I include compiling beta drivers into Un*x kernels in that time. When I was part of running the ISP, Macs were <10% of our customer base, and >80% of our tech support time. Your experience may have been good, but in the course of an evening I've watched a massage-therapist Mac friend go from crashing every 5 minutes, to reinstalling the OS and running just fine for half an hour, to installing IE5 again, to crashing every 5 minutes, to reinstalling the OS and swearing never to touch IE again. I've had similar experiences where I did the "poke through the extensions folder and remove everything that look MSish and it worked again. Obviously, our respective mileages varied.

Colors: Yep, I tried to recreate the original default Netscape link colors (which is what I thought I'd had at one point, but I guess I'd screwed up). I'm not happy with the white on black colors, but I'll try to actually do a coherent design once I get back from Burning Man and get the new colo box in place.

Re: 3.0, 3.2, 2.0: I'm actually wondering now if it was 2.0 and not 3.0, I suppose I could go dig through the source and see what it assumes when it can't find a doc type. Damn my faulty memory. But in general my beef with the Web Standards Project is that they'll issue some grand proclamation, and then spend the next few months individually saying "but that's not what we really meant" on message boards, mailing lists, and private email. If I get the people involved cornered individually I've generally found them to be quite reasonable (although I don't think I'd have that experience with Zeldman), but as a whole I see the announcements as generally poorly thought out and inflammatory in a way that doesn't provoke good discussion.

And I do need to put doc type headers (HTML4 loose) on the Flutterby pages, and actually double-check and make sure that they're all XHTML too. Now if I can just coordinate with my favorite sites on CSS names so that I can go CSS in a way that lets users customize more than one site at a time I might actually add a cookie-based mode that runs as CSS.

#Comment made: 2001-08-25 02:35:54+00 by: John Anderson [edit history]

> if I can just coordinate with my favorite sites on CSS names so that I can
> go CSS in a way that lets users customize more than one site at a time I
> might actually add a cookie-based mode that runs as CSS.

I'd be up for that. Actually, the way to go might be to try to define a Standard Blog Element Sytle Sheet (TM) and then just publish it. You wouldn't need *that* many elements -- title, subtitle, date, accent color, link colors, etc...

PS: Why am I not getting quote display when I start a line with '>'?

#Comment made: 2002-02-21 05:32:34+00 by: Pete

You know that 4.7 is up to 4.78, right? They even managed to get something right that had bugged me for a looong time--separate toggles for javascript on the web and usenet.

#Comment made: 2002-02-21 05:32:35+00 by: Dan Lyke

John: No clue, that should work just fine. Are you using Mac IE?

Pete: Heck, Opera is up to 5, and completely rocks. The point is that I don't always have the option of installing new browsers on all the machines I work on. Heck, I have enough problems getting a recent version of Perl sometimes.

#Comment made: 2001-08-26 03:02:56+00 by: Pete [edit history]

These go to eleven.

I don't have a position I'm espousing. I just thought the info could be of interest to people still monkeying around with the Netscape 4 series.

I tried Opera years back, but it was a mess then, and it's subwindow approach was a turnoff for me.

#Comment made: 2001-08-26 11:12:44+00 by: John Anderson [edit history]

Dan: No, I'm using Opera 5.05beta under Linux. <shrug> I guess if it really matters to me, I'll start doing HTML posts with BLOCKQUOTE...

Hey, wasn't there some talk about getting some sort of threading going in here? Most of the time it would be overkill, but for these longer discussions, it would be really nice.

#Comment made: 2002-02-21 05:32:35+00 by: Larry Burton

Maybe an option to sort these comments in ascending or descending order?

The standardized naming convention for weblog stylesheets would be something that would really interest me. As stated it wouldn't need that many classes and IDs but just having standard names would sure make me feel better about creating the stylesheets. I'd rather my creativity, what little there may be, go to something other than thinking up names for the different elements of the document.