CSS on demand
January 28, 2003 9:17 AM Subscribe
CSS on Demand allows users to set several preferences for how they want to see your site, rather than just using one of your themes via a switcher. Kind of like Matt lets you do here.
Perl. Free. Try it out.
Perl. Free. Try it out.
Well, that was informative.
As for the broken link, I dunno what happened. This one works.
posted by Su at 10:03 AM on January 28, 2003
As for the broken link, I dunno what happened. This one works.
posted by Su at 10:03 AM on January 28, 2003
I nice tool. I would like to see how this could be implemented with something other than a blog.
posted by psycht at 11:05 AM on January 28, 2003
posted by psycht at 11:05 AM on January 28, 2003
pretty cool, simple and elegant solution (and the links seem to be working at the moment), but definately not CSS.
of course CSS itself, when correctly implemented, allows for user selection via the title attribute of the style element.
posted by danOstuporStar at 11:29 AM on January 28, 2003
of course CSS itself, when correctly implemented, allows for user selection via the title attribute of the style element.
posted by danOstuporStar at 11:29 AM on January 28, 2003
Mod note: You can also use ID on your BODY tags to let users customize your entire site with their style sheets. via diveintomark
posted by kirkaracha (staff) at 12:08 PM on January 28, 2003
posted by kirkaracha (staff) at 12:08 PM on January 28, 2003
interesting little script... but I really don't know if it accomplishes much more then pointing out how flawed current browsers are when it comes to implementation & management of user stylesheets. (on preview, same with the method discussed on css-d)
posted by 10sball at 12:14 PM on January 28, 2003
posted by 10sball at 12:14 PM on January 28, 2003
10sball: Without spending a lot of time attaching styles in a specific manner to a site(because style, id, etc. names are not standardized), user stylesheets will only let you do really basic things like choose a base font.
I can see your point, but I think that this idea could be taken even further than that, if the concept is expanded. Your comment limits it to being nothing more than a stylesheet switcher, which isn't the case at all; it's a customizer. Problem is that it just doesn't do too much more than a style switcher would at the moment.
Think about a site with a base layout(think 2x2 grid a la FilePile*), the user could specify which topic went in which quadrant, by having this trade out positioning values. There are sites that have functionality somewhat like this, but it was coded from the ground up, and likely with much more involved tools than just a stylesheet. Something like this would be a nice utility. It would require some serious expansion, but I can dream, and this thing's only at v0.1, anyway.
*Original layout. I haven't seen the site in ages.
posted by Su at 12:38 PM on January 28, 2003
I can see your point, but I think that this idea could be taken even further than that, if the concept is expanded. Your comment limits it to being nothing more than a stylesheet switcher, which isn't the case at all; it's a customizer. Problem is that it just doesn't do too much more than a style switcher would at the moment.
Think about a site with a base layout(think 2x2 grid a la FilePile*), the user could specify which topic went in which quadrant, by having this trade out positioning values. There are sites that have functionality somewhat like this, but it was coded from the ground up, and likely with much more involved tools than just a stylesheet. Something like this would be a nice utility. It would require some serious expansion, but I can dream, and this thing's only at v0.1, anyway.
*Original layout. I haven't seen the site in ages.
posted by Su at 12:38 PM on January 28, 2003
By the same token, Su, I think you may be short changing what /could/ be possible if a browser truly let a user manage local stylesheets.
Here's one example that I've used elsewhere in the past. It's m more like the css-d example but in browser and not reliant on authors consent to using ids. Perhaps a "Make new stylesheet" function that would load the current document in an editor like window with a little toolbar that would record all the changes you made to the display of the document - both simple things like changing font sizes & potentially more complicated things like positioning. You could then hit the little save button and have the option to "save for all sites" "save for this site" or "save for this page only" and it would then remember to toss the newly generated CSS doc into the right spot in the cascade whenever you revisited the site. So yes, under my newfangled enhanced stylesheet management you could reorder the boxes of a 2x2 grid.
Even then, I can see how expansion of this reprocessing concept is interesting, but if you're really dreaming why not just make the jump to a magical time and place where every site validates as some type of xml based markup and you can just run local xslt transformations to *really* flow things however you'd like? [one can always dream]
posted by 10sball at 1:06 PM on January 28, 2003
Here's one example that I've used elsewhere in the past. It's m more like the css-d example but in browser and not reliant on authors consent to using ids. Perhaps a "Make new stylesheet" function that would load the current document in an editor like window with a little toolbar that would record all the changes you made to the display of the document - both simple things like changing font sizes & potentially more complicated things like positioning. You could then hit the little save button and have the option to "save for all sites" "save for this site" or "save for this page only" and it would then remember to toss the newly generated CSS doc into the right spot in the cascade whenever you revisited the site. So yes, under my newfangled enhanced stylesheet management you could reorder the boxes of a 2x2 grid.
Even then, I can see how expansion of this reprocessing concept is interesting, but if you're really dreaming why not just make the jump to a magical time and place where every site validates as some type of xml based markup and you can just run local xslt transformations to *really* flow things however you'd like? [one can always dream]
posted by 10sball at 1:06 PM on January 28, 2003
Or one could just use Opera. Which, if I'm not mistaken, just released the version 7 beta today.
posted by five fresh fish at 1:08 PM on January 28, 2003
posted by five fresh fish at 1:08 PM on January 28, 2003
I gotta point out, though, that you've got to be one helluva geek to run about making custom CSS.
posted by five fresh fish at 1:09 PM on January 28, 2003
posted by five fresh fish at 1:09 PM on January 28, 2003
Moz, Opera, and IE all let you write a single user style sheet that's on all the time, useful if your eyesight is bad, or you're just really anal about having control over how sites look, but not much more.
If Opera7 (which i think went gold and not beta) gives you more control (e.g. turn it on or off based on site you're on) well then kudos to them, but I wouldn't know cause I'm on a mac :/
posted by 10sball at 1:16 PM on January 28, 2003
If Opera7 (which i think went gold and not beta) gives you more control (e.g. turn it on or off based on site you're on) well then kudos to them, but I wouldn't know cause I'm on a mac :/
posted by 10sball at 1:16 PM on January 28, 2003
but if you're really dreaming why not just make the jump to a magical time and place where every site validates as some type of xml based markup and you can just run local xslt transformations to *really* flow things however you'd like?
You are, in both of your examples, giving too much work to the user, though your visual idea is nice. But that's a backwards approach. I want a development tool. The site designer will use it, and the user just gets a series of buttons saying "I want this font", and "Put this there," etc.
I'll risk saying that my suggestions still fall well within the realm of possibility, now, as opposed to "in some magical time", whereas to respond to your suggestions, it's still hard to find a site that validates as even HTML, much less any X(HT)ML variant. And how many people are actually using XSLT right now, other than in an enterprise environment?
As for the CSS side of it, anybody that would be interested in using something like this to develop their site will likely be doing a decent job of it, and if not, then it'd probably be broken in the default view, in the first place.
It's not incorrect to complain that the browsers don't make good use of user/alternate stylesheets, but it's also practical to accept that it's probably not going to be fixed anytime soon(alt. sheets have been around since CSS1, and are only actually showing up in only some browsers now), and look for another way.
posted by Su at 1:29 PM on January 28, 2003
You are, in both of your examples, giving too much work to the user, though your visual idea is nice. But that's a backwards approach. I want a development tool. The site designer will use it, and the user just gets a series of buttons saying "I want this font", and "Put this there," etc.
I'll risk saying that my suggestions still fall well within the realm of possibility, now, as opposed to "in some magical time", whereas to respond to your suggestions, it's still hard to find a site that validates as even HTML, much less any X(HT)ML variant. And how many people are actually using XSLT right now, other than in an enterprise environment?
As for the CSS side of it, anybody that would be interested in using something like this to develop their site will likely be doing a decent job of it, and if not, then it'd probably be broken in the default view, in the first place.
It's not incorrect to complain that the browsers don't make good use of user/alternate stylesheets, but it's also practical to accept that it's probably not going to be fixed anytime soon(alt. sheets have been around since CSS1, and are only actually showing up in only some browsers now), and look for another way.
posted by Su at 1:29 PM on January 28, 2003
ahh... but its no fun to spend time contemplating reality!
Let me just end my involvement by saying that its been my experience as a surfer that the pages I most want to alter in some way are those exactly that *wouldn't* think enough to offer these tools server side. So the problem, as I approach it, has got to be solved on the client side.
posted by 10sball at 1:38 PM on January 28, 2003
Let me just end my involvement by saying that its been my experience as a surfer that the pages I most want to alter in some way are those exactly that *wouldn't* think enough to offer these tools server side. So the problem, as I approach it, has got to be solved on the client side.
posted by 10sball at 1:38 PM on January 28, 2003
Even then, I can see how expansion of this reprocessing concept is interesting, but if you're really dreaming why not just make the jump to a magical time and place where every site validates as some type of xml based markup and you can just run local xslt transformations to *really* flow things however you'd like? [one can always dream]
Question: how difficult would it be to write an automated script to turn function HTML into valid XHTML? I would be willing to bet that at least half of current sites could be made valid by closing BR and IMG tags, closing unclosed tags at the last valid point (i.e., immediately before the containing tag is closed), adding a doctype declaration, adding quotes around attribute values, and a few other minor changes, at which point an arbitrary XSLT stylesheet could be applied. I'm thinking it would be fairly trivial, for example (although it's been a few months since I've messed with XSLT, so I may be mistaken) to write a stylesheet that would show MetaFilter minus posts with < n comments, or posts by certain users, or in alphabetical order or what have you. The advantage of this over PHP or Perl is that you wouldn't be dependent on the site's author to implement many features.
Of course, the challenge is to come up with a way to do this that doesn't require the end user to learn XSLT. Evolved algorithms trained by positive and negative examples, anyone? More practically, present the user with something like the DOM viewer in Mozilla, or IE's default stylesheet for XML, and allow them to select particular elements and base rules on them. (Sort by this element, hide this element, hide element A when X attrib of element B is foo, etc.)
Well, ok, maybe 'trivial' isn't quite the right word. But it should be possible with current technology and current web coding practices.
posted by IshmaelGraves at 5:29 PM on January 30, 2003
Question: how difficult would it be to write an automated script to turn function HTML into valid XHTML? I would be willing to bet that at least half of current sites could be made valid by closing BR and IMG tags, closing unclosed tags at the last valid point (i.e., immediately before the containing tag is closed), adding a doctype declaration, adding quotes around attribute values, and a few other minor changes, at which point an arbitrary XSLT stylesheet could be applied. I'm thinking it would be fairly trivial, for example (although it's been a few months since I've messed with XSLT, so I may be mistaken) to write a stylesheet that would show MetaFilter minus posts with < n comments, or posts by certain users, or in alphabetical order or what have you. The advantage of this over PHP or Perl is that you wouldn't be dependent on the site's author to implement many features.
Of course, the challenge is to come up with a way to do this that doesn't require the end user to learn XSLT. Evolved algorithms trained by positive and negative examples, anyone? More practically, present the user with something like the DOM viewer in Mozilla, or IE's default stylesheet for XML, and allow them to select particular elements and base rules on them. (Sort by this element, hide this element, hide element A when X attrib of element B is foo, etc.)
Well, ok, maybe 'trivial' isn't quite the right word. But it should be possible with current technology and current web coding practices.
posted by IshmaelGraves at 5:29 PM on January 30, 2003
« Older Challenger Explodes | Doomed planet found Newer »
This thread has been archived and is closed to new comments
posted by BobInce at 9:37 AM on January 28, 2003