I'm considering two changes to the stats pages, and I'd like your input.

  1. CSS - I have a version of the pages redone in CSS. There's no real change in the presentation, but it does make the HTML more W3C compiant (thanks to AgentM for running a scanner, and noticing these). The majority of browsers seem to support CSS, but I know that AgentM is running iCab, and it doesn't support CSS. How many people would this effect, and what's your opinion on changing over to that?

  2. Cookies - Over the months, there are a few people who are rather outspoken on their distaste for cookies. However, if I added (non-mandatory) cookie support, I could allow setting a default page, and perhaps a few other features. I don't want to add user/login support to keep track of preferences. Would non-mandatory cookies be OK with you folks?
--Chris

e-mail jcwren

Replies are listed 'Best First'.
(jcwren) Re: Stats change proposal question
by jcwren (Prior) on Jan 29, 2001 at 19:43 UTC
      Unfortunately, Netscape 4.7x (Windows) renders your table cells with a black background.

      There may be a couple of tricks that might help.

      • Consider embedding CSS into the content instead of using an external stylesheet (which shouldn't be too difficult as you're using PHP).

        (Update: ichimunki raises a good point. Sadly, Netscape 4.7 has a bug regarding linked stylesheets. This is only one reason why I occasionally rant about non-compliant browsers.)

      • Make sure you define rules for <TD>, <TH>, and <TR>.

      • Make sure you use and close your <P> tags.

      Note sure if you've seen this site, but it contains a lot of CSS related information, including a page documenting many of the most egregious problems.

      Hope this helps...

      --f

      P.S. As far using CSS and cookies go, I'm with Blue: Go for it.

        My own bias would be against embedding style information into the HTML document unless absolutely necessary to assist the user agent in displaying the information (i.e. layout is integral to the information being displayed). Usually this is not the case. Putting the CSS into its own document makes the original document smaller, and heightens awareness that style is not content on the part of the designer.
        Actually, If you EVER close a P tag, you need to close them all, I've had better results never closing them. Also, I've learned not to mix inline style with external stylesheets since you can't ever tell what will happen when you do. All major browsers seem to have one issue or another with mixing inline and external.

        --
        $you = new YOU;
        honk() if $you->love(perl)

Re: Stats change proposal question
by Corion (Patriarch) on Jan 29, 2001 at 20:41 UTC

    Under Konqueror 2.1 beta2, the new stats page looks good.

    If you try doing CSS with Netscape, you might want to consider the two tricks I know about CSS and Netscape :

    1. Netscape CSS JavaScript reload : When resizing a NetScape browser window, all CSS information gets lost from time to time. The ugly fix is to have a JavaScript onresize handler, which reloads the complete page after each resize.
    2. You can manage to load different CSSheets for IE and Netscape by using two or more <link rel="..." id="foo1"> links and one link without the id attribute. Netscape will be oblivious to all link tags with the id attribute, while you can select under IE (via nasty JavaScript) the "correct" CSS by setting the enabled attribute to "1" (if I remember right).

Re: Stats change proposal question
by Fastolfe (Vicar) on Jan 29, 2001 at 21:43 UTC
    1. As much as I love strict HTML4.1-compliant pages with CSS to do the formatting, you'll probably have to go with the masses here. Lots of people are (for whatever reason, legitimate or not) using browsers incapable of handling standards released several years before the end of the last millennium. Some might argue that this is the chicken-and-the-egg thing. Without a reason to clammor for support of these standards, they'll never be supported. Whenever I'm writing something (non-work-related, obviously), I do tend to push things into CSS-land, but I will add a trivial tag or two here and there (thus making my HTML "transitional" rather than "strict") to at least make it look "OK" in non-CSS browsers. They will never look nearly as good though, and I think that's OK. It's just enough of an incentive.

    If it's just a vote you're looking for, I vote for CSS.

    2. I still don't understand the anti-cookieism attitude. Most browsers have options such as "accept cookie only from originating site" and I think Mozilla even has a mechanism to accept cookies only from certain hosts. If you want to give into the evil tracking paranoia (which I will admit isn't 100% unfounded), you do have options here. Non-mandatory cookies seems like a nice compromise. If people want to live without them, they just won't see those nifty features. I think that's fine. Everyone is happy that way.

Re: Stats change proposal question
by lemming (Priest) on Jan 29, 2001 at 20:08 UTC

    Looks like Netscape 4.7 on windows doesn't pick up the nwp class so it looks like black on black. So I would suggest for anyone back there to upgrade to 4.75 which looks fine. Better in fact than it looks under the old way.

    IE 5.0 & 5.5 look fine.

    I'm actually in favor of CSS, now that I've looked at it. As long as you don't get carriedf away with some of the features, since even the browsers that are CSS compliant are really only 60% functional. And you can get some wierd results with borders, DIV, and floating elements.

    On cookies: Non-mandatory cookies sounds good to me.

    Update: Just read your rants about netscape. You're right, IE does do better than NS. They seem to fail in different spots, IMO NS will fail in some of the more used areas. From what I've read Netscape 6 is supposed to be very compliant to CSS, but I haven't bothered with it yet.

    YAU: Hmm. The black on black is because of the javascript needs to be on to use CSS properly under netscape.

Re: Stats change proposal question
by TStanley (Canon) on Jan 29, 2001 at 20:27 UTC
    CSS - Looks great for me on Netscape at work, and I run IE 5.5 at home,so that
    shouldn't be a problem.

    Cookies: Non-mandatory cookies sound good to me. At least you are giving the person
    an option to use or not to use.

    TStanley
    In the end, there can be only one!
Re: Stats change proposal question
by AgentM (Curate) on Jan 29, 2001 at 21:57 UTC
    Personally, I think your pages look great as they are now! :-D It doesn't seem anymore that anyone pays attention to those W3C kooks. Anyway, I understand your shift to standards compliance and i was glad to help you but sadly, I do see some demons.

    The aforementioned Netscape bug is especially brutal but that JavaScript insert is a wasteful, unneccessary hack. "Let's dare to be JavaScript free." (Right, Petruchio?)

    While cookies are an option, another simple thing you may want to consider is simple CGI browser recognition where you send me and Mozilla 4.7 folks to the non-CSS page and the rest (or those that can handle it) to the "new" page. Frankly, I don't see a whole lot of benefit from cookies unless you've thought of something impressive that I haven't. Honestly, at least in my iCab, your page looks great right now as it is. I'd prefer to use that despite its HTML pitfalls rather than some JScript hacked, variable CSS, cookie monster.

    AgentM Systems nor Nasca Enterprises nor Bone::Easy nor Macperl is responsible for the comments made by AgentM. Remember, you can build any logical system with NOR.
Re: Stats change proposal question
by Blue (Hermit) on Jan 29, 2001 at 19:52 UTC
    I'm fine with CSS. Love them personally. I've foudn that cutting down on font tags and the like that get repeated often (such as in cells of tables) cuts down my load time, too.

    I think most of the currect browsers will support non-inline CSS (excect anything AgentM runs {grin}).

    As for cookies, when I don't have them disabled I have them pop-up a window, which is annoying. But the functionality you mention would be nice. So as long as it doesn't try to re-set them every page, I'm fine with them. You're adding additional functions for people who don't mind them, adn not taking away from the people who don't. That's fair.

    =Blue
    ...you might be eaten by a grue...

Re: Stats change proposal question
by stefan k (Curate) on Jan 29, 2001 at 20:55 UTC
    Dear Monks,

    CSS: the main problem with CSS-enhanced sites is that I don't see them properly because f****ing Netscape (4.7) simply drops CSS if JavaScript isn't enabled. Maybe it would be possible to set this in my user-settings? Having a box saying "Use CSS" sounds great to me.

    Cookies: As I'm used to clicking away mere loads of cookie warnings I'd say yes to cookies. Yet again: it would be nice to have it user-defined. Or you could set a cookie which decides whether the visitor accepts them or not. It is _very important_ that I don't have any cookie windows whenever I hit the Reload-Button on the Newest Nodes page, which appears to happen quite often (at least 3 times an hour on a usual work-day)
    Regards
    Stefan K

    $dom = "skamphausen.de"; ## May The Open Source Be With You! $Mail = "mail@$dom; $Url = "http://www.$dom";
Re: Stats change proposal question
by ichimunki (Priest) on Jan 29, 2001 at 20:50 UTC
    I think removing all layout related markup from any web page (except for the DIV and SPAN tags, which also serve to mark logical blocks as well as provide hooks for style sheets) is always a good idea. While many browsers do not handle the fanciest features of CSS well or consistently, uncoupled style makes it more likely that your page will render well in browsers other than Netscape and IE. It also allows users the option in those browsers to easily override your styles.

    IIRC, the stats pages are already well done, and the move to CSS should be smooth.

    Personally, I have no objection to cookies as long as I can choose to not accept them and still get to all of the information.
Re: Stats change proposal question
by elusion (Curate) on Jan 30, 2001 at 01:35 UTC
    I like CSS. Most browsers people use these days support them. Except for people with really old browsers or non-mainstream ones. I think people need to start using them. CSS has been around for a while, and lots of people haven't used them because old browsers don't support them. It's time that we force people to upgrade.

    But what I was going to say... It looks great in Mozilla.

    - p u n k k i d
    "Reality is merely an illusion, albeit a very persistent one." -Albert Einstein

Re: Stats change proposal question
by Maclir (Curate) on Jan 30, 2001 at 02:19 UTC
    My views:
    1. CSS - go for it. NetScrape 6 is not quite as broken as NutScape 4, but the main table still has no border, and there are different font faces and sizes on the "news" page between IE5.5 and NS6. I personally prefer the smaller, serif (Times New Roman?) that I see with NS 6 than the larger sans-serif font IE5.5 used.

      As well, using the closing paragraph </p> is good practice - XML will eventually force such things.

    2. Cookies - give a reason why not, people. If you don't like them, turn them off. You then lose the ability to have your preferances saved. Tough.

    My only other comment on the pages is could you give a thought on the news page to having dates in the format "27 Jan 2001" rather than "01/27/2001"? I know it is only us non-americans that are radical and prefer to see the day before the month, but spelling out the month name means it is obvious. Then you will just have the non-english speakers complaining about the use of the english abbreviations. Sigh. Maybe use the same date format as the date / time stamps on the main page ("2001-01-29")