Microsoft in trouble yet again....

Inane baanter or words of wisdom - you decide...
Post Reply
User avatar
Dio
Sheepie Whore
Sheepie Whore
Posts: 18887
Joined: Mon May 31, 2004 1:08 pm 0
Location: Newcastle UK

Re: Microsoft in trouble yet again....

Post by Dio »

I highly doubt it's that little work.

But even beyond then, it's the principle of the thing: target a standard version and hit it. It's like there's no point in making a video card that can do DX9 with one thing from DX10 - it's all or no point.
User avatar
oNyx
Baaaaa!
Baaaaa!
Posts: 1652
Joined: Mon Sep 08, 2003 1:24 am 0
Location: de/fd

Re: Microsoft in trouble yet again....

Post by oNyx »

Well, you gotta register the property, parse it, and if it exists you draw some rounded rect instead of a rect. Doesn't sound all that complicated to be honest - especially if you consider that step 1+2 are basically already done.
User avatar
Dio
Sheepie Whore
Sheepie Whore
Posts: 18887
Joined: Mon May 31, 2004 1:08 pm 0
Location: Newcastle UK

Re: Microsoft in trouble yet again....

Post by Dio »

What about the interactions with all the other styles? Isn't that the whole point of the C in CSS?

And even then, you've missed out unit tests and regression tests - absolutely essential with something like a browser. What's the cost of reviewing 100 million web pages to check it hasn't broken anything?
User avatar
xav
Baaaaaa!
Baaaaaa!
Posts: 2455
Joined: Thu Jun 06, 2002 5:32 pm 0
Location: Bucks, UK
Contact:

Re: Microsoft in trouble yet again....

Post by xav »

oNyx wrote:Well, you gotta register the property, parse it, and if it exists you draw some rounded rect instead of a rect. Doesn't sound all that complicated to be honest - especially if you consider that step 1+2 are basically already done.
It may not sound that complex, but there's perhaps more to it than meets the eye:

* The spec details edge cases - such as border-radius being applied to rects with different border thicknesses. If the standard Windows "draw a rounded rect" primitive doesn't handle those edge cases in the same way, there's likely to be more work involved than you suggest.

* Rounded corners also present problems if the content doesn't fit inside the rounded box. For example, a square-cornered image in a rounded div. Firefox's early implementation of rounded corners suffered from this - the image would poke out of the div at the corners - but I think this have been fixed now. Heck, there was a time when even applying a background colour in Firefox would cause the corners to poke out of rounded borders.

* It's possible (though this is pure speculation) that they decided to defer implementation of border-radius until they've got the time to go all the way with border-image as well (due to be supported in Firefox 3.1)

* Plus, as Dio pointed out, it makes sense for them to draw a line under CSS2 and stop there. There are a whole load of useful CSS3 things that would seem to be fairly easy to implement, but if they don't stop somewhere, IE8 will never see the light of day.


Don't get me wrong, I'd love to see more widespread border-radius support - but I'd rather live without it to get IE8 out sooner. The sooner it's out, the sooner web developers will be able to target CSS2.1 as a base spec.
Co-creator of The Greys, Monsters, Inked & Elvie webcomics
• Web Site • Twitter • Patreon •
User avatar
xav
Baaaaaa!
Baaaaaa!
Posts: 2455
Joined: Thu Jun 06, 2002 5:32 pm 0
Location: Bucks, UK
Contact:

Re: Microsoft in trouble yet again....

Post by xav »

Mitchell Baker has posted about this. She's the "Chief Lizard Wrangler" at Mozilla - i.e. the woman in charge. She makes some interesting points - especially about whether or not the rise of Firefox represents "healthy" competition in the browser space:

http://blog.lizardwrangler.com/2009/02/ ... microsoft/
Co-creator of The Greys, Monsters, Inked & Elvie webcomics
• Web Site • Twitter • Patreon •
User avatar
Little stoned kiwi
Baaaaaa!
Baaaaaa!
Posts: 4598
Joined: Mon Oct 27, 2003 6:05 am 0
Location: Christchurch

Re: Microsoft in trouble yet again....

Post by Little stoned kiwi »

Did you all hear the hoopla about MS's latest .NET patch? It installs the .NET CLickOnce add-on to Firefox, without declaring it in the text, without asking for permission, and without a removal facility.

If you can't beat them, infiltrate them.
User avatar
gumboot
Baaaaaaa!
Baaaaaaa!
Posts: 7652
Joined: Sun Aug 26, 2007 3:42 am 0
Location: trumpstain!

Re: Microsoft in trouble yet again....

Post by gumboot »

She puts me on Microsoft's side in just the second paragraph...
the damage Microsoft has done to [...], and the pace of the web development itself is both glaring and ongoing.
Granted the arms race at the start did nothing but harm, but when they turned around and started dragging their heels they nobbled web developers' ability to adopt new standards. That may be a downer for web developers, but it saves the end user from the perpetual cycle of upgrades that don't actually make anything work any better but do introduce a stream of new bugs over the top of old.
User avatar
gravy
Sheepie Whore
Sheepie Whore
Posts: 19572
Joined: Wed May 01, 2002 1:28 pm 0
Location: The Other Place

Re: Microsoft in trouble yet again....

Post by gravy »

Dio wrote:What about the interactions with all the other styles? Isn't that the whole point of the C in CSS?

And even then, you've missed out unit tests and regression tests - absolutely essential with something like a browser. What's the cost of reviewing 100 million web pages to check it hasn't broken anything?
It's only going to affect pages with the border radius tag in. Which isn't any where near a 100 million pages.

I agree it's a bit of non trivial work that'll need thorough testing but that it probably won't be in an MS browser until the far off IE9 isn't great. MS's browser dev seems to be slow and expensive and leads to the weakest end product of all the major browser suppliers.
User avatar
xav
Baaaaaa!
Baaaaaa!
Posts: 2455
Joined: Thu Jun 06, 2002 5:32 pm 0
Location: Bucks, UK
Contact:

Re: Microsoft in trouble yet again....

Post by xav »

gumboot wrote:She puts me on Microsoft's side in just the second paragraph...
the damage Microsoft has done to [...], and the pace of the web development itself is both glaring and ongoing.
Granted the arms race at the start did nothing but harm, but when they turned around and started dragging their heels they nobbled web developers' ability to adopt new standards. That may be a downer for web developers, but it saves the end user from the perpetual cycle of upgrades that don't actually make anything work any better but do introduce a stream of new bugs over the top of old.
So by letting IE stagnate and disbanding the development team Microsoft were actually protecting users. If only they'd stagnated with a few of their other products, we could all be running DOS and Word 1.0 safe in the knowledge that they weren't introducing any new bugs.

As for the "upgrades that don't actually make anything work any better", I would argue that the only reason why the web doesn't seem to work any better than it did a few years ago is because Microsoft keep dragging their heels about adding more modern features. If it wasn't for the need to support IE, websites would already "work better":
  • They would work faster - every other vendor has put a lot of effort into optimising their javascript engines. Microsoft hasn't.
  • They would look prettier - every other browser has supported transparent PNGs for years. Microsoft only added proper support in IE7.
  • They would be more powerful - things like <canvas>, which lets a developer draw arbitrary shapes and images to the screen. Supported in every major browser except IE.
  • They would have vector graphics - basic SVG support is present in all the other browsers. Good enough for drawing graphs and diagrams at least. Not supported at all in IE.
  • They would be developed faster, cheaper - IE's lack of standards support adds significant extra development time and cost.
Without the huge numbers of IE users adding an effective barrier to the use of modern standards, the web would be a far more interesting and powerful development platform than it is right now.

IE8 is due out this year, finally adding comprehensive support for CSS2 - a standard that was finished in 1998! That's what web developers are having to contend with - developing for a platform which doesn't even meet the standards of a decade ago. Hurrah for that protective stagnation.
Co-creator of The Greys, Monsters, Inked & Elvie webcomics
• Web Site • Twitter • Patreon •
User avatar
gumboot
Baaaaaaa!
Baaaaaaa!
Posts: 7652
Joined: Sun Aug 26, 2007 3:42 am 0
Location: trumpstain!

Re: Microsoft in trouble yet again....

Post by gumboot »

xav wrote:
gumboot wrote:She puts me on Microsoft's side in just the second paragraph...
the damage Microsoft has done to [...], and the pace of the web development itself is both glaring and ongoing.
Granted the arms race at the start did nothing but harm, but when they turned around and started dragging their heels they nobbled web developers' ability to adopt new standards. That may be a downer for web developers, but it saves the end user from the perpetual cycle of upgrades that don't actually make anything work any better but do introduce a stream of new bugs over the top of old.
So by letting IE stagnate and disbanding the development team Microsoft were actually protecting users. If only they'd stagnated with a few of their other products, we could all be running DOS and Word 1.0 safe in the knowledge that they weren't introducing any new bugs.
I don't care about the version of the product. I care about the version of the standard. Encouraging Windows users to use a shitty browser isn't my problem. The benefit comes from encouraging webdevs to stop jumping on more modern standards that require browser development to focus on new features rather than finishing off the features that already exist.

xav wrote:Without the huge numbers of IE users adding an effective barrier to the use of modern standards, the web would be a far more interesting and powerful development platform than it is right now.
And this is precisely my point. I don't give a fuck what it's like as a development platform. It's a protocol for pushing information about, and I want webdevs to focus on the content and orderly presentation rather than wanking over all the fucking bling they'd have if only they could make everybody use the most recent standard.
User avatar
Ratsoap
Baaaaaaa!
Baaaaaaa!
Posts: 6344
Joined: Fri Dec 19, 2003 1:46 pm 0
Location: Back in Edinburgh
Contact:

Re: Microsoft in trouble yet again....

Post by Ratsoap »

gumboot wrote: And this is precisely my point. I don't give a fuck what it's like as a development platform. It's a protocol for pushing information about, and I want webdevs to focus on the content and orderly presentation rather than wanking over all the fucking bling they'd have if only they could make everybody use the most recent standard.
Content = HTML/XML, presentation = CSS.

CSS in IE = Broken.

We're not trying to set the world on fire, it would just be nice to be able to work to the standards that are out there, without having to waste time and money on cack IE.

Sometimes it's just nice to vent, man.
Because fucking marketing cunts would never call a spade anything other than a Deluxe Easi-Shift Earth Movement Solution.
User avatar
xav
Baaaaaa!
Baaaaaa!
Posts: 2455
Joined: Thu Jun 06, 2002 5:32 pm 0
Location: Bucks, UK
Contact:

Re: Microsoft in trouble yet again....

Post by xav »

gumboot wrote:The benefit comes from encouraging webdevs to stop jumping on more modern standards that require browser development to focus on new features rather than finishing off the features that already exist.
Just out of interest, what existing features do you consider to be unfinished in browsers?
Co-creator of The Greys, Monsters, Inked & Elvie webcomics
• Web Site • Twitter • Patreon •
User avatar
gumboot
Baaaaaaa!
Baaaaaaa!
Posts: 7652
Joined: Sun Aug 26, 2007 3:42 am 0
Location: trumpstain!

Re: Microsoft in trouble yet again....

Post by gumboot »

Ratsoap wrote:Content = HTML/XML, presentation = CSS.
CSS in IE = Broken.

We're not trying to set the world on fire, it would just be nice to be able to work to the standards that are out there, without having to waste time and money on cack IE.
Well, to be honest, I have distinct memories of the web working just fine before CSS came about. Of course, they were the days of "best viewed in 1280x1024" pages, and shit like that, but that's not something that CSS itself actually solved... and there are still plenty of pillarbox websites that are made that way by their style sheets.

xav wrote:
gumboot wrote:The benefit comes from encouraging webdevs to stop jumping on more modern standards that require browser development to focus on new features rather than finishing off the features that already exist.
Just out of interest, what existing features do you consider to be unfinished in browsers?
It depends on the browser, but in general terms the "not crashing" feature and the "going fast" feature are in desperate need of work. Some browsers (normally the ones that have all the go-fast and not-crash) lack CSS or Javascript altogether.
User avatar
xav
Baaaaaa!
Baaaaaa!
Posts: 2455
Joined: Thu Jun 06, 2002 5:32 pm 0
Location: Bucks, UK
Contact:

Re: Microsoft in trouble yet again....

Post by xav »

gumboot wrote:
Ratsoap wrote:Content = HTML/XML, presentation = CSS.
CSS in IE = Broken.

We're not trying to set the world on fire, it would just be nice to be able to work to the standards that are out there, without having to waste time and money on cack IE.
Well, to be honest, I have distinct memories of the web working just fine before CSS came about. Of course, they were the days of "best viewed in 1280x1024" pages, and shit like that, but that's not something that CSS itself actually solved... and there are still plenty of pillarbox websites that are made that way by their style sheets.
That depends on your definition of "working just fine" I suppose. CSS didn't really add many styling capabilities that weren't already there. It was already possible to make an abomination of a web page using <font> tags and bgcolor attributes. CSS simply separated much of that out to a separate file - and gave the end user the ability to override the author's choice of styling if they so wished. Not that many people know about that last bit. That's the one area in which browsers fail with CSS: by not making it easier for the user to edit their local stylesheet.

Without CSS there would be no View=>Page Style=>No Style menu option in Firefox - and I would have no real recourse when faced with a web page with some godawful combination of colours and fonts. Heck, it even works on pillarbox websites made that way by their style sheets.

gumboot wrote:
xav wrote:Just out of interest, what existing features do you consider to be unfinished in browsers?
It depends on the browser, but in general terms the "not crashing" feature and the "going fast" feature are in desperate need of work. Some browsers (normally the ones that have all the go-fast and not-crash) lack CSS or Javascript altogether.
I guess I'm just lucky, but I've never really been afflicted by the crashing feature at all. As for the "going fast" feature...

You claim that having IE development stagnate was good because it prevented the developers being distracted by flashy new features so they could concentrate on fixing the current issues. You then point to "going fast" as one of the current issues. But given that IE is considerably slower than its competitors, how exactly do you think that stagnation has helped to fix that issue? While the other, non-stagnant browsers have been working on new, faster rendering code and javascript engines, stagnant IE has only barely improved in speed.


Have you considered that bringing IE up to par might actually make web pages load faster:
  • Every line of hacky IE workaround that gets removed reduces the download size and processing burden
  • PNGs could be more widely used - they're usually smaller than equivalent GIF images
  • SVG could be used, where appropriate, to include (potentially interactive) graphs and diagrams with a small file size
  • Rounded corners in CSS are far more lightweight than the image, CSS, code and markup alternatives that prevail
  • Multiple columns defined in CSS avoid the need for table markup or javascript solutions whilst being more lightweight and flexible
Not every new feature necessarily implies bloat and slowdown. Some of the newer CSS effects are aimed at moving common tasks away from the slow and heavy Javascript+images approaches that are used today, towards implementation in the core of the browser - where they can be made fast and low-impact.
Co-creator of The Greys, Monsters, Inked & Elvie webcomics
• Web Site • Twitter • Patreon •
User avatar
Dio
Sheepie Whore
Sheepie Whore
Posts: 18887
Joined: Mon May 31, 2004 1:08 pm 0
Location: Newcastle UK

Re: Microsoft in trouble yet again....

Post by Dio »

We've already had the rounded corners argument: if they were that important, they would have been in CSS2.
Post Reply