On Netscape 4.x and the Stylesheets of LJ's S2 Layout Engine:

[acroyear]

Now I don't want to get into a rant here, but neither IE NOR Netscape follow the standards correctly. I support the standards, and also respect that one must move forwards in capabilities or be left behind by those that do. The browser wars were supposed to be about standards. Tragically, Netscape 1) lost the war and 2) didn't have a stable-enough code base to be fixed via open source methods (particularly when required 3rd party stuff was removed, leaving a lot of gaping functionality holes).

Thus, Mozilla had to basically start over w/ the new layout engine, Gecko. Newer *IS* better in this case, both in the Mac world and in the standards-based browsing world.

The style sheets are NOT poorly formated. They are perfectly legit to the standard; i ran 'em through a validator. That Netscape for doesn't (and will never) support the standard is neither our fault, nor LJs. That Netscape freezes or crashes as a result is certainly not our fault.

One can always try to go into the netscape configuration options dialog and turn off stylesheets. Netscape 4 allowed that, at least for windows/linux, though the resultant look was often horid and even more unreadable. The other option, seeing if the right-mouse-button menu had a "copy link location" (which it does in IE and Mozilla) isn't an option for OS 9 since Macs don't have that right mouse button to start with (and thus, most mac apps don't have those menus even for the macs with 2-button mice).

Mozilla as "newest and greatest" it most certainly isn't. Html 4.0 and CSS 2.0 are *3 years old*. Netscape 4.x is technically now 5 years old and 5 years abandoned. No bug fixes or functionality fixes have been done in ANY Netscape 4 release since 4.6. The only reason ANY release after 4.6 happened is to fix security issues to keep the feds (who'd committed to Netscape for a solution back in the 3.x days) from complaining. That's it. No other bugs like the multitude of CSS problems have ever been fixed in a Netscape 4 release since 4.6 over 5 years ago.

So no, I have no intention any more of crippling my stuff to work with a product that has been abandoned by its owners for half a decade.

The world moves on.

Comments

[acroyear70.livejournal.com]

crawl because somebody says I should turn on JavaScript on an underpowered machine

javascript is not (and never was) java; java is a JVM that when it launches, takes between 4 and 10 seconds and once running can still be dog slow on an underpowered box.

but javascript is just a little scripting language that is built-in to the browser and is "there" even when turned off. i know how its implemented in several browsers and aside from the "pop-ups" problem (where some adverts pop up and then bring up more popup ads as they get killed and you have all the memory wasted by the pages and the slowdown of the loading and all that rot) it really doesn't do anything intensive. the delays in most image rollover scripts are caused by loading the images, not in the rollover code itself.

i know you mentioned javascript being a slowdown before and it confused me then, too.

i do know that on netscape 4, having css on and javascript off, or vice-versa, broke a lot of apps/pages because dHTML assumes that both are there -- each is dependent on something done by the other.

---

I know that some browsers claim some degree of CSS compliance that isn't there. Obviously if something as simple as a CSS file causes a browser to freeze or crash, there's a big problem (and not one I can or should have to debug on my end).

For uglyness problems, that's where different browsers supported different parts of the standard first, all in the promise (that few actually ever met; IE still doesn't and likely won't for that matter) of fixing the rest later. positioning is usually the big one, like where netscape 4 ignores page margins on tables but nothing else (it means that table-controlled layout from netscape 2|3 gets screwed up with border images).

failing "gracefully" is relative. from a browser's perspective: once its tried to layout something, and then fails to layout something else, what do you do? re-render the whole page without the stylesheet? can you even detect that you failed to lay it out right? if you ignore some css entries and its not a problem, but others and it is, how can you tell the difference?

the "promise" of CSS being backwards-compatible and graceful in its failures was bogus, and I wrote so on a web-bb back in 1998 after having played with netscape 4 for a few months, Then i had to go to considerable work to make my pages look good for both css and not. some parts of celticdistrict.com are still broken and really just need a total redesign that i lack the patience for right now.

Finally, I didn't write the stylesheet for my blog; it was a built-in one to the particular s2 layout I used. my blog does exist for my comfort first.

maybe we should file a bug report/feature request that "style=mine" become an option that can be permanent and affect all pages, whether its in the link or not. there are other usability/accessability reasons besides just lousy-CSS that people would prefer to have their "look" be what they use for the whole site.

Some people intentionally use larger fonts because of eye-sight problems; having huge fonts on their own page, then having to enlarge the fonts (if their browser supports it) every time they read a lj-cut or comment, then shrink them back down when they get back to their own/friends page is a tiresome excercise too.

---

Firefox is still a memory hog, don't get me wrong on that. I was using Mozilla 1.4 on my last underpowered laptop (p2-233, 32meg memory, win98) and pretty much had to turn all other apps off to come close to being usable. But it was still stable and didn't give any page-rendering problems except for those pages whose javascript assumed that mozilla 5 == mozilla 4 and made bad DHTML requests.

Opera 7 is probably the more efficient as far as memory goes, but i can't confirm that.

end part 1. part 2 coming up.

[dglenn.livejournal.com]

Java is more of a dog than Javascript, yes. But here's the rub: I can leave Java turned on because very few pages use it -- I get a slow launch, performance is dragged down somewhat for a little while, I go elsewhere, performance is fine again.

Javascript, whether it's supposed to be painless or not, is everywhere, so if I turn it on, I've got script after script all trying to run at once. And when a page uses Javascript, it tends to use a lot of it. The theory is interesting, but I find my direct observations more interesting: when I turn on Javascript, within fifteen minutes my machine is so bogged down that I'm waiting thirty to a hundred and twenty seconds for response to a keypress or mouseclick, and it doesn't get better until I turn Javascript off.

The problem is not pop-up windows; I keep those disabled.

Until the arrival of the 350 MHz equipment, the fastest machines I ran browsers on were:

Win95, Pentium, 120 MHz, I think 48 MB RAM.
WinNT, Pentium, 100 MHz, 80 MB RAM.
Mac, have to look up CPU and speed, 128 MB RAM.

Whether Javascript is supposed to bog those machines down or not is secondary to the fact that it does. "Just a little scripting language" that gets used for bigger and bigger scripts. (In many cases I feel it's The Wrong Tool, but that's a different discussion.)

"Finally, I didn't write the stylesheet for my blog; it was a built-in one to the particular s2 layout I used. my blog does exist for my comfort first."

*nod* It's true that specific complaints about specific bugs relating to specific S2 styles should be addressed to the LJ support staff.

And I've already acknowledged that your priorities regarding accessibility vs. style differ from mine. My personal decision is to favour a style less likely to cause problems, but that wasn't my complaint about what you had written. My complaint was with the "Oh just upgrade already" tone.

(BTW, your current LJ style does not break under Opera 6. The problems I've encountered personally have been elsewhere, and some of those are still readable but I can't imagine the rendering I'm seeing is intentional, while others either have foreground and background the same colour or completely omit the text until I tack on "?style=mine".)

'maybe we should file a bug report/feature request that "style=mine" become an option that can be permanent and affect all pages, whether its in the link or not.'

Good idea. Make it part of the login cookie.

Thanks for the caveat regarding Firefox.

Turning off all other apps isn't a good solution for me. I keep less stuff active at a time than I'd really like already; at some point I wind up spending more time closing and restarting apps than I do using them.