OPERA GENERAL 35 RE ALTERNATE STYLESHEET PERSISTANCE SUCKS BADLY
From: Kevin W (kevin_w@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sat, 12 Jul 2003 11:25:39 +1000


> I was playing with adding a new alternate stylesheet on my blog but as > far as I can tell they are mostly useless.
> If I reload the page or visit another page then my choice of stylesheet > is forgotten and it goes back to the default one.
>
> Am I missing something?

No you're not. Opera (and Mozilla and Konqueror) don't remember the selected stylesheet between reloads or different pages. Yes, it sucks.

-- Kevin W :-)
Opera/CSS/webdev blog: http://trats.ozforces.com.au/
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/












From: Headless (invalid_address@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sat, 12 Jul 2003 08:26:24 +0100

Paul McGarry <paul@no-spam> wrote:

>I was playing with adding a new alternate stylesheet on my blog but as far >as I can tell they are mostly useless.
>If I reload the page or visit another page then my choice of stylesheet is >forgotten and it goes back to the default one.
>
>Am I missing something?

Normal behaviour (every UA does it like that). Imo it would be wrong to assume that if a user selects an alternate stylesheet on one page, he would want that to apply to other pages as well, so the way that UA's currently do it is correct imo. (besides, other pages may not offer the same, any, or different alternate stylesheets)

If you want an alternate stylesheet to be persistent across a site then you need to implement a server side solution, or if that isn't possible use a client side solution.

I use a js client side solution:
http://alistapart.com/stories/alternate/
This one actually has a potential advantage over server side solutions in that by default it also works if a user uses for example Opera's menu to switch to an alternate stylesheet (normally you place a stylesheet switching widget on the site itself). It uses document.unload and a cookie to achieve that, drawback is that this causes a cookie to be set even if the user doesn't use an alternate stylesheet. I considered that to much of a drawback, so on my sites I've modified it so that it only sets a cookie if a user elects to use an alternate stylesheet.

Rijk recently mentioned another (client side?) method:
>The switcher at http://devedge.netscape.com/
Headless

From: Headless (invalid_address@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sat, 12 Jul 2003 08:43:13 +0100

Headless <invalid_address@no-spam> wrote:

>I use a js client side solution:
>http://alistapart.com/stories/alternate/
>
>Rijk recently mentioned another (client side?) method:
>>The switcher at http://devedge.netscape.com/
Correction, Devedge also used the js client solution that I'm using,
only their (also modified) version of it doesn't work for me (7.11).

Headless

From: Lauri Raittila (lauri@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sat, 12 Jul 2003 10:46:06 +0300

In article <a5cvgvcgsjo2alkelfh94995atj11s86lo@no-spam>, Headless wrote:
> Paul McGarry <paul@no-spam> wrote:

> >Am I missing something?

No, unfortunately > Normal behaviour (every UA does it like that). Imo it would be wrong to > assume that if a user selects an alternate stylesheet on one page, he > would want that to apply to other pages as well, so the way that UA's > currently do it is correct imo. (besides, other pages may not offer the > same, any, or different alternate stylesheets)

Do you really think so yourself? Would be sorry if it would stick?

I would say it would be odd to assume he don't want to use the alternative style for other pages. In fact, this is first time I ever hear anybody say s/he would prefer the current behaviour.

-- Lauri Raittila <http://www.iki.fi/lr> <http://www.iki.fi/zwak/fonts>
Saapi lähettää meiliä, jos aihe ei liity ryhmään, tai on yksityinen tjsp., mutta älä lähetä samaa viestiä meilitse ja ryhmään.


From: Headless (invalid_address@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sat, 12 Jul 2003 16:37:11 +0100

Samuli Lintula <none@no-spam> wrote:

>> Imo it would be wrong >> to assume that if a user selects an alternate stylesheet on one page,
>> he would want that to apply to other pages as well, so the way that >> UA's currently do it is correct imo.
>
>You could make the same case for User mode (Ctrl+g), couldn't you? Why >shouldn't a user selection stick?

Ctrl-G pertains to the browser window, selecting an alternate stylesheet pertains to the document being viewed.

Headless

From: Headless (invalid_address@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sat, 12 Jul 2003 16:40:34 +0100

Lauri Raittila <lauri@no-spam> wrote:

>> Normal behaviour (every UA does it like that). Imo it would be wrong to >> assume that if a user selects an alternate stylesheet on one page, he >> would want that to apply to other pages as well, so the way that UA's >> currently do it is correct imo. (besides, other pages may not offer the >> same, any, or different alternate stylesheets)
>
>Do you really think so yourself?

Duh.

>Would be sorry if it would stick?

I believe it would be fundamentally wrong to do that, apart from the principle it could also screw up existing solutions and it would make it far more difficult to create browser independent solutions.

Headless

From: Lauri Raittila (lauri@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sat, 12 Jul 2003 21:28:22 +0300

In article <hpa0hvc26i9mn3fqv06136iucu6qm7hgbq@no-spam>, Headless wrote:
> Lauri Raittila <lauri@no-spam> wrote:

> >Would be sorry if it would stick?

> I believe it would be fundamentally wrong to do that,

I am very surprised. HTML spec is more than little unclear and CSS spec is no good either. There is need for better specs IMO.
But it would be absolutely wrong to not allow UA to have sticking alternative stylesheets.

> apart from the > principle it could also screw up existing solutions and it would make it > far more difficult to create browser independent solutions.

Why would it broke existing solutions? The problem with all current solutions is that they aren't very good. For example, I can't assign shortcut key to alternative styles using current, author side solutions. (Already I can assign one in Opera - but it is still so buggy it will crash browser instead of toggling alternative stylesheet.)

-- Lauri Raittila <http://www.iki.fi/lr> <http://www.iki.fi/zwak/fonts>
Saapi lähettää meiliä, jos aihe ei liity ryhmään, tai on yksityinen tjsp., mutta älä lähetä samaa viestiä meilitse ja ryhmään.


From: Tom Gilder (tom@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sat, 12 Jul 2003 21:47:23 +0100

Paul McGarry wrote:
> I was playing with adding a new alternate stylesheet on my blog but as > far as I can tell they are mostly useless.
> If I reload the page or visit another page then my choice of stylesheet > is forgotten and it goes back to the default one.

I'm still awaiting a browser to make text sizing, user style sheets, alternative style sheets, and various other settings actually useful by keeping a database of what setting is applied to what site.

If it can be done for cookies and passwords, why not other settings?


From: Headless (invalid_address@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sat, 12 Jul 2003 23:05:53 +0100

Lauri Raittila <lauri@no-spam> wrote:

>> >Would be sorry if it would stick?
>
>> I believe it would be fundamentally wrong to do that,
>
>I am very surprised. HTML spec is more than little unclear and CSS spec >is no good either. There is need for better specs IMO.

The html spec is perfectly clear on this:

>The author may specify that one of the alternates is a preferred style >sheet. User agents should apply the author's preferred style sheet >unless the user has selected a different alternate.

Alternate stylesheets belong to a _single_ document only. It may well be that all documents on a site are equipped with the same set of alternate stylesheets, but this should not be assumed by a UA.

It's perfectly valid and correct to have documents on the same site with identically named but completely different alternate stylesheets. A UA that tries to make alternate stylesheets persistent site wide could cause the wrong stylesheet to be selected. It also clearly goes against what the html spec says (see quote above).

Only the site author knows if it is appropriate to make identically named alternate stylesheets persistent site wide, thus it is for the author to create a solution for this, not something a UA should attempt to guess.

>> apart from the >> principle it could also screw up existing solutions and it would make it >> far more difficult to create browser independent solutions.
>
>Why would it broke existing solutions?

A UA attempting to make an alternate stylesheet persistent would end up fighting existing client side solutions trying to do the same thing.

>The problem with all current >solutions is that they aren't very good. For example, I can't assign >shortcut key to alternative styles using current, author side solutions.
There are 2 types of author side solutions of making alternate stylesheets persistent site wide:

1) Pure server side selection of another stylesheet. This method doesn't list any alternate stylesheets in the html.
2) By using client side scripting, this does list alternate stylesheets in html, thus they can be selected via the UA's UI, and thus they can have a hotkey assigned to them.

Headless

From: Headless (invalid_address@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sat, 12 Jul 2003 23:05:54 +0100

Samuli Lintula <none@no-spam> wrote:

>>>You could make the same case for User mode (Ctrl+g), couldn't you? Why >>>shouldn't a user selection stick?
>> >> Ctrl-G pertains to the browser window, selecting an alternate >> stylesheet pertains to the document being viewed.
>
>That's a very unclear distinction, in my opinion. And a very >author-centric view, to say the least. If we look at the situation from >a user's point of view (the Opera way), then in both cases a user >chooses a way to get a "better Internet experience" and that is what >Opera should give them.
That's a very short sighted view. These things need to be thought through whilst considering the ramifications that go beyond such simplistic thoughts.

Headless

From: Paul McGarry (paul@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sun, 13 Jul 2003 10:19:09 +1000

On Sat, 12 Jul 2003 23:05:53 +0100, Headless <invalid_address@no-spam> wrote:

> The html spec is perfectly clear on this:
>
>> The author may specify that one of the alternates is a preferred style >> sheet. User agents should apply the author's preferred style sheet >> unless the user has selected a different alternate.
>
> Alternate stylesheets belong to a _single_ document only.

So you keep saying, nothing in that quote indicates a users selection applies to a single document only.

> It may well be > that all documents on a site are equipped with the same set of alternate > stylesheets, but this should not be assumed by a UA.

The UA doesn't need to assume anything.
It needs to keep using the same style sheet if possible and go back to the default if not.

> It's perfectly valid and correct to have documents on the same site with > identically named but completely different alternate stylesheets.

a) Of course it's possible for a web author to display bad usability traits such as giving vastly different style sheets the same name.
b) Having different style sheets with the same name isn't necessarily bad. If there was a "High visability" named alternate and one page needed a bit of special tweaking to achieve the desired effect having a different stylesheet (in terms of URI) that has the same effect.

> A UA that tries to make alternate stylesheets persistent site wide could > cause the wrong stylesheet to be selected.

Whereas now if I've selected an alternate style sheet and gone to another page it _always_ causes the wrong stylesheet to be loaded.

> It also clearly goes against what the html spec says (see quote above).

Stop saying that.

> Only the site author knows if it is appropriate to make identically > named alternate stylesheets persistent site wide
A stylesheets name is there to indicate to the user what the stylesheet is likely to do.
As a user I would assume that two stylesheets with the same name are likely to have a highly similar effect.
As such I see know reason why the user agent shouldn't be able to do the same on my behalf.

I see no reason to hobble the alternate stylesheet functionality in Opera out of fear that authors can't label their stylesheets sensibly.

, thus it is for the > author to create a solution for this, not something a UA should attempt > to guess.

It's a user agent. It should do what the user wants it to do.

> A UA attempting to make an alternate stylesheet persistent would end up > fighting existing client side solutions trying to do the same thing.

Who cares? AFAICT in 99% of cases any "fight" would involve both solutions trying to do the same thing.
The fact that "existing solutions" have been created because the alternate stylesheet behaviour in browsers has been woefully lacking is not a reason not to fix it in the right place.

-- Paul McGarry http://pmcg.blogspot.com/


From: Samuli Lintula (none@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: 13 Jul 2003 08:38:03 GMT

Headless <invalid_address@no-spam> wrote in news:k611hvkjign1mjmuolmc933kt3fo73tr08@no-spam
> Samuli Lintula <none@no-spam> wrote:
>
>>That's a very unclear distinction, in my opinion. And a very >>author-centric view, to say the least. If we look at the situation >>from a user's point of view (the Opera way), then in both cases a user >>chooses a way to get a "better Internet experience" and that is what >>Opera should give them. > > That's a very short sighted view. These things need to be thought > through whilst considering the ramifications that go beyond such > simplistic thoughts.

Right. But where is your argument?

-- Samuli Lintula

From: John Lewis (gleemax@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sun, 13 Jul 2003 04:59:22 -0500

On Sat, 12 Jul 2003 23:05:53 +0100, Headless <invalid_address@no-spam> wrote:

> The html spec is perfectly clear on this:
>
>> The author may specify that one of the alternates is a preferred style >> sheet. User agents should apply the author's preferred style sheet >> unless the user has selected a different alternate.

Yes, it's perfectly clear that there's no mention of documents.

> Alternate stylesheets belong to a _single_ document only.

No. Alternate style sheets belong to all documents that link to them.

> It may well be > that all documents on a site are equipped with the same set of alternate > stylesheets, but this should not be assumed by a UA.

No. If document A and document B link to the same style sheet(s), Opera will know. If document B doesn't link to the alternate style sheet selected in document A, then obviously that style sheet cannot be chosen. You're free to express your opinion, but you're mistaken if you think it's impossible or that HTML says it shouldn't be done for the good of mankind.

> It's perfectly valid and correct to have documents on the same site with > identically named but completely different alternate stylesheets.

No. Two different style sheets don't have the same URI.

> A UA attempting to make an alternate stylesheet persistent would end up > fighting existing client side solutions trying to do the same thing.

Only if the author intends for that to happen (or is amazingly incompetent) .

-- John Lewis, using the helpful W3-dev menu for Opera 7 through the courtesy of Toby Inkster:
<http://goddamn.co.uk/~tai/?id=10>


From: Eirik Mikkelsen (eir-mik@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sun, 13 Jul 2003 11:59:31 +0200

On Sat, 12 Jul 2003 18:59:45 +1000, Paul McGarry <paul@no-spam> wrote:

> On Sat, 12 Jul 2003 08:26:24 +0100, Headless <invalid_address@no-spam> > wrote:
>
>> Normal behaviour (every UA does it like that). Imo it would be wrong to >> assume that if a user selects an alternate stylesheet on one page, he >> would want that to apply to other pages as well, so the way that UA's >> currently do it is correct imo.
>
> It's bonkers. It's equally "wrong" to assume that a user _doesn't_ want > to use it on others.
>
> If a subsequent page has an alternate stylesheet with the same name and > uri as a previous page it seems a no brainer to use that one.
>
>> (besides, other pages may not offer the >> same, any, or different alternate stylesheets)
>
> If other pages don't offer the same stylesheets then obviously it should > revert to the default sheet.
>
>
>

Completely agree with you Paul! This has been bugging me too.
And to make everybody happy this setting should be a toggle in Preferences- >Page style
-- Eirik

From: Headless (invalid_address@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sun, 13 Jul 2003 13:10:43 +0100

John Lewis <gleemax@no-spam> wrote:

>> Alternate stylesheets belong to a _single_ document only.
>
>No. Alternate style sheets belong to all documents that link to them.

Declaring a stylesheet alternate is done by setting the rel attribute to "alternate stylesheet" and by giving it a title. For a UA the title is the way to distinguish between them. It should be possible for an author to use 2 different stylesheets with an identical title. A UA attempting to make an alternate stylesheet persistent across a site takes away that valid and correct option, and it conflicts the w3c rule that an alternate stylesheet should not be applied until the user has chosen to do so. This pertains to a single document only. Authors have the option to specify different behaviour.

Headless

From: Headless (invalid_address@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sun, 13 Jul 2003 13:12:02 +0100

Samuli Lintula <none@no-spam> wrote:

>Right. But where is your argument?

Right in front of you.

Headless

From: Paul McGarry (paul@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sun, 13 Jul 2003 22:41:34 +1000

On Sun, 13 Jul 2003 13:10:43 +0100, Headless <invalid_address@no-spam> wrote:

> and it conflicts the w3c rule that an alternate stylesheet should not be > applied until the user has chosen to do so.

> This pertains to a single document only.

Stop saying that, it's total bollocks!
The spec doesn't say it. It simply does not define by which mechanisms a user can choose an alternate stylesheet.

The fact is when I go to a page and choose an alternate stylesheet I have chosen that stylesheet for subsequent pages.
The problem is that the mechanisms currently available for expressing that choice are not sufficient.

There are lots of people who want a better solution and the lack of decent implementation guidelines is acknowledged by "top" CSS people.

http://lists.w3.org/Archives/Public/www-style/2003Jun/0067.html
-- Paul McGarry http://pmcg.blogspot.com/


From: John Lewis (gleemax@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sun, 13 Jul 2003 08:25:27 -0500

On Sun, 13 Jul 2003 13:10:43 +0100, Headless <invalid_address@no-spam> wrote:

> John Lewis <gleemax@no-spam> wrote:
>> [Headless wrote:]
>>> Alternate stylesheets belong to a _single_ document only.
>>
>> No. Alternate style sheets belong to all documents that link to them.
>
> Declaring a stylesheet alternate is done by setting the rel attribute to > "alternate stylesheet" and by giving it a title.

Yes. That doesn't conflict with what I said; in fact, it's part of the foundation for my argument, and runs counter to your argument. The title is not used to retrieve the style sheet any more than the URI is used to name the style sheet. Referencing a style sheet in the link element is done with a URI, as I already mentioned in my previous response.

I've already addressed this in my other message. You responded to a fraction of my response with an argument already addressed in another part of my response. Don't bother replying to this message if you're going to do the same thing again; you'll be wasting my time as well as your own.

> For a UA the title is the way to distinguish between them.

No, the title is for the user to distinguish between style sheets. For the UA it only determines which type of style sheet the style sheet is interpreted as. With alternate style sheets, it's only used by the UA for grouping on a document-level basis, which is entirely irrelevant to this discussion.

> It should be possible for an author > to use 2 different stylesheets with an identical title.

It is, and would remain possible. Why wouldn't it be? The value of the title attribute is wholly unrelated to the location/contents of the style sheet.

-- John Lewis, using the helpful W3-dev menu for Opera 7 through the courtesy of Toby Inkster:
<http://goddamn.co.uk/~tai/?id=10>


From: Samuli Lintula (none@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: 13 Jul 2003 13:46:00 GMT

Headless <invalid_address@no-spam> wrote in news:63j2hvgsnar45g4lvi08n3te5mnjuelrta@no-spam
> Samuli Lintula <none@no-spam> wrote:
> >>Right. But where is your argument?
> > Right in front of you.

In your latest message, your argument was basically that I'm shallow and stupid. For an argument, ad hominems aren't very good.

-- Samuli Lintula

From: Lauri Raittila (lauri@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sun, 13 Jul 2003 17:11:42 +0300

In article <opn0hvo8ooo0krqeb3ddpn5mikesnkh1ba@no-spam>, Headless wrote:
> Lauri Raittila <lauri@no-spam> wrote:
> > >> >Would be sorry if it would stick?
> >
> >> I believe it would be fundamentally wrong to do that,
> >
> >I am very surprised. HTML spec is more than little unclear and CSS spec > >is no good either. There is need for better specs IMO.
> > The html spec is perfectly clear on this:
> > >The author may specify that one of the alternates is a preferred style > >sheet. User agents should apply the author's preferred style sheet > >unless the user has selected a different alternate.

And where it mentions more than one document? Especially, where it mentions that this procedure is not to be used on multible documents for one user selection? Spec just don't specify what to do on case we are talking. And, that is afaik because no-one ahd any idea about current situation when HTML spec was written.
But that is not the point. There it is not forbidden to have sticky stylesheets, just because it is not mentioned in spec. Showing target URL in status bar is not mentioned in spec either (afaik). I think you wouldn't like it it was not shown somewhere?

It is very short sighted to deny this just because you have nice little stylesheet swicher that works on all new browsers, in some settings. That is only argument against sticky stylesheets we have heard from you. (exept for claim it is not according to spec, but it's very questionable if spec addresses this at all)

I can't have shortcuts. I can't have any better UI.

And who says your swicher wouldn't work afterwards? I see no reason why your style swicher wouldn't work, if someone uses build-in version. If it would break, it does it currently too, when someone swiches stylesheet using browser UI. Or am I mistaken?

-- Lauri Raittila <http://www.iki.fi/lr> <http://www.iki.fi/zwak/fonts>
Saapi lähettää meiliä, jos aihe ei liity ryhmään, tai on yksityinen tjsp., mutta älä lähetä samaa viestiä meilitse ja ryhmään.


From: Manni Heumann (manni.heumann@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: 13 Jul 2003 15:42:20 GMT

Coises wrote:

> Unfortunately, as far as I know there is no unambiguous,
> algorithmic definition of what constitutes a single site.

No, there probably isn't. But this is beside the point. If I have two documents, referencing the same style sheet (that is, the same URI *and* the same title) and the user has selected that same style sheet as his preferred style sheet, the browser may safely assume, that the user wants that style sheet again, when it sees it referenced in some page.
This scheme is totally unrelated to sites, hosts, or even domains. Here's an example:

document A at host1.domain.tld/something lists a style sheet at www.domain.tld/styles/big.css titled "Big".
document B at host2.domain.tld/whoknows lists the same style sheet with the same name.

All the browser has to do is to look in his (I guess rather short) list of selected alternate style sheets that contains URIs and names and say "Hey, my master has selected this one before, he is the boss, so I will give it to him again".

Note that this would even work with document C at www.different-
domain.tld/xy.

If somebody specifies an alternate style sheet in one of his documents, we should simply assume that it will work, no matter where his document or that style sheet can be found.

Manni
PS: And if some people cannot live with that because (add some reason) here, OS should do what has always helped in the past: Add an option like "[ ] Remember my choice of alternate styles". No reason for any fights.


From: Paul McGarry (paul@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Mon, 14 Jul 2003 01:54:37 +1000

On Sun, 13 Jul 2003 07:47:39 -0700, Coises <Opera-spamtrap@no-spam> wrote:

> There is a possibly troublesome situation here.
>
> Suppose we have:
>
> www.big.isp/members/user1/page1.htm > <link rel="alternate stylesheet" type="text/css" title="big" > href="b.css">
>
> www.big.isp/members/user2/start.htm > <link rel="alternate stylesheet" type="text/css" title="big" > href="b.css">
>
> Here we wouldn't want a selection of "big" for the first document to > carry forward to the second.

While that might not be ideal, I don't think it would be the end of the world if the UA did stick with the "big" style selection accross these two "sites". Assuming that authors title their stylesheets reasonably then a user who selects the "big" stylesheet on one of these sites is probably not going to have a major issue with a stylesheet of the same title being applied on the other.

It's also worth pointing that occasionally serving up the "wrong" stylesheet on a site as in this example is still an improvement on the current situation where the "wrong" stylesheet is served up every time because the UA simply forgets the users choice.

There is certainly going to be room for tweaking the algorithm but these problems aren't going to be fixed by sitting around theorising about them. An implementation needs to be made with fairly simple rules and we need to see how that works in the real world.

> It does seem to me that selections of alternate stylesheets should > persist > throughout a site. Within a site, it would be reasonable to expect that > the same alternate stylesheet name (the "title" attribute value) would > refer to the same *intent*, regardless of whether the author uses the > same > or different CSS files to achieve it; and it's the name, not the URI, > that > users see, and would expect to persist.

I agree 100%. The URI should play no part in it. (I think I may have said otherwise earlier)

> Unfortunately, as far as I know there is no unambiguous, algorithmic > definition of what constitutes a single site. That would force a browser > to go by URI --- which, I've tried to show above, is bound to generate > inconsistencies from the user's perspective, since users think in terms > of "sites" and alternate stylesheet names, not stylesheet URIs (which > they normally don't even see).

It doesn't "force the browser to use the URI".
I'm sure using the URI is the wrong way to go. I think it adds complexity and doesn't really promote clarity.

Instead we can just accept that in some rare edge cases it might result in a user getting an alternate stylesheet where they might otherwise the default one and that even if this happens it is unlikely to be harmful.

I expect that a relatively simple solution will present itself to authors worrying about such "persistance spillover".

It may be as simple as including something like: <link rel=start href="http://www.big.isp/members/user1/">

(using "start" might be a misuse but another word could work, rel=root or something).

> My point is that the situation is ambiguous enough that we can't say > Headless doesn't have a point (even though I think alternate stylesheet > selections *should* apply throughout a site, if a way to make that work > as expected can be devised).

Currently it works 100% unexpectedly. Almost anything would be an improvement.
There are no guidelines in CSS in part because no-one has implementation experience and they don't want to put something in the spec that is practically unfeasible.

Someone needs to step up with a reasonable implementation and see how it works in practice and then we can tweak from there.

In reality I expect cases where -2 sites at the same domain have alternate stylesheets of the same title.
-it will be harmful for the user to be presented with such an alternate stylesheet at site B after choosing it at site A.
are incredibly rare.

-- Paul McGarry http://pmcg.blogspot.com/


From: Headless (invalid_address@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sun, 13 Jul 2003 20:23:41 +0100

John Lewis <gleemax@no-spam> wrote:

>> It may well be >> that all documents on a site are equipped with the same set of alternate >> stylesheets, but this should not be assumed by a UA.
>
>No. If document A and document B link to the same style sheet(s), Opera >will know.

A function like that could be created, but it would only move the problem somewhere else, this is another perfectly valid and correct construct:

<link rel="stylesheet" title="Preferred" href="preferred.css"
type="text/css" />
<link rel="alternate stylesheet" title="Alternate" href="alt1.css"
type="text/css" />
<link rel="alternate stylesheet" title="Alternate" href="alt2.css"
type="text/css" />

From the html spec:

>Authors may group several alternate style sheets (including the author's >preferred style sheets) under a single style name. When a user selects a >named style, the user agent must apply all style sheets with that name.
Your proposed solution of saving the css file location in some way and then use that in combination with the title to identify an alternate stylesheet would wreck that type of construct.

There are perfectly good solutions if an author wants to make an alternate stylesheet stick across a site (cookie or session variable),
it's not a task for a UA.

Headless

From: Paul McGarry (paul@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Mon, 14 Jul 2003 07:59:59 +1000

On Sun, 13 Jul 2003 20:23:41 +0100, Headless <invalid_address@no-spam> wrote:

> There are perfectly good solutions if an author wants to make an > alternate stylesheet stick across a site (cookie or session variable),
> it's not a task for a UA.

Complete rubbish!

CSS is an independant technology and alternate stylesheets should be fully functional without relying on:
- Cookies - Javascript - Server side voodoo.

Authors should not have to donate a part of their web page to style switching.

A user should have a consistant user interface for altering style. If a user has accessibility problems with the currently selected style they should just be able to pop open the style menu to choose another one.

-- Paul McGarry http://pmcg.blogspot.com/


From: John Lewis (gleemax@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Sun, 13 Jul 2003 18:53:16 -0500

On Sun, 13 Jul 2003 20:23:41 +0100, Headless <invalid_address@no-spam> wrote:

> Your proposed solution of saving the css file location in some way and > then use that in combination with the title to identify an alternate > stylesheet would wreck that type of construct.

Style sheet grouping is done on a document level, independently of any style sheet URIs. Try again.

-- John Lewis, using the helpful W3-dev menu for Opera 7 through the courtesy of Toby Inkster:
<http://goddamn.co.uk/~tai/?id=10>


From: Tom Gilder (tom@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Mon, 14 Jul 2003 02:42:31 +0100

Headless wrote:
> There are perfectly good solutions if an author wants to make an > alternate stylesheet stick across a site (cookie or session variable),
> it's not a task for a UA.

And a webpage could provide an input for going to a new URL, instead of the browser having an address bar.

But no, hang on, then the address bar might be in a different place on every site. And the address bar might act differently.

If the browser can do something which is helpful to the user in a predictable way without breaking any specification then there's nothing wrong with it. At all.


From: Kevin W (kevin_w@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Mon, 14 Jul 2003 12:09:11 +1000

I agree completely with all of your points, Paul!

It is not the end of the world if the UA chooses an undesirable stylesheet, as long as it stays within the bounds of the spec. The user can always change it back, and it certainly beats having to select the desired style over and over and over.

(almost) Any implementation other than the current one will be an improvement. A simple one could be to check that the pages have the same *set of titles*.

If persistence namespace spillover becomes a (very minor) problem, then authors can easily fix it by changing their stylesheet titles to "meyerweb.Eos" or whatever.

If the UA is in doubt at any point over which stylesheet to choose, just apply the preferred stylesheet as it normally would.

-- Kevin W :-)
Opera/CSS/webdev blog: http://trats.ozforces.com.au/
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/


From: Kevin W (kevin_w@no-spam)
Subject: Re: Alternate stylesheet persistance sucks badly
Date: Mon, 14 Jul 2003 12:12:18 +1000

> There are perfectly good solutions if an author wants to make an > alternate stylesheet stick across a site (cookie or session variable),
> it's not a task for a UA.

If that were the case, then I wouldn't need the Back button in my browser. Just tell all of the webmasters to add a JavaScript link to go Back on all of their pages.

-- Kevin W :-)
Opera/CSS/webdev blog: http://trats.ozforces.com.au/
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/