Posts in the Tech Category

My first suggestion for improving the W3C is this:  every Working Group should have one member whose primary (and possibly sole) responsibility is outreach.

To make life a little easier, I’m going to refer to this position as a WGO (for Working Group Outreach).  As an aside, I’m not sure that “outreach” is exactly the right term for what I have in mind, but it’s a decent term that captures most of what I have in mind, so I’ll use it here.  If someone comes up with a better term, I’ll be grateful.

So here’s what I envision for a WGO.

1. The WGO keeps the public informed about the top issues on the Working Group’s agenda and immediate-future activities.  The easiest, most obvious way to do this is to post a summary of every WG FTF (face-to-face) meeting.  A summary would describe the topics the WG discussed, resolutions that were reached, which problems were not solved, and so forth.  This could be a bullet-point list, but a better summary would be something like a short article.

   Note that I do not say that the WGO should post the FTF minutes, which are often private.  The results of those discussions, though, should be public, even when no results occurred.  A summary can say that the WG discussed a topic at length and reached no resolution without saying why.  It can also say that a topic was discussed and a solution found, and then describe the solution.

   A really good WGO would produce an activity summary more often than every FTF.  I don’t know that I’d insist on a summary for every weekly teleconference, but sending out a summary once a month would be more than reasonable.  These summaries would be posted on the W3C site and to the relevant public mailing lists.  For the CSS WGO, this would always mean posting to www-style.  In cases where WG activity touched on features of XHTML or SVG, summary posts would be made to those public lists as well.

   The purpose here is to draw back some of the curtain surrounding Working Groups.  Too often, interested members of the public don’t know what the WG is up to, and that can be frustrating.  If there are several people agitating for a new feature and the WG stays silent on it, it’s impossible to tell if the WG is blowing the idea off or if it’s something they’ve considered at length but haven’t yet reached a decision.

   Public summaries also have the benefit of allowing some public discussion of work before the public-comment period on a proposed specification.  This would help distribute the WG’s feedback load.

2. The WGO brings the needs and concerns of the public to the Working Group, and communicates back the WG’s reactions.  This means part of the WGO’s job is to be involved in the wider community surrounding a given activity.  The CSS WGO, for example, would spend time reading web design mailing lists, forums, blogs, and so forth to find out what people in the field want and need (in CSS terms, anyway).  The WGO would present these to the WG as items to consider.  The topics so raised, and the WG’s responses to them, would go into the next summary.

   The goal here, of course, is to have someone on the Working Group who represents the “in the trenches” folks.  If there are other members of the WG who also represent those who work in the field, that’s awesome.  With the WGO position, though, there’s the assurance of at least one person who speaks for those who actually use the products of the Working Group, and who will use any future products.

3. The presence of a WGO in a Working Group should be a charter condition.  No group should be (re-)chartered without an identified WGO, and the extended lack of a WGO should be cause to question the continued charter of a group.

   Basically, I’m of the opinion that if a WG can’t find someone passionate enough about what they’re doing to be the WGO, then it’s time to ask whether or not they should continue at all.  Similarly, if there’s no real community for the WGO to represent, then it’s time to ask why the WG even exists.

4. The WGO should have no other major responsibilites within the Working Group.  This means the WGO cannot be the WG’s chair, and should not be a specification editor.  Their primary job should be the two-way representation I’ve described here.

   It’s too easy to get overloaded in a WG, especially if you’re the kind of enthusiast a good WGO should be.  There needs to be a defined limit to the position, so that outreach is always topmost on that person’s agenda within the WG, and it doesn’t get buried under other duties.

In summary, a good WGO would act as a liason between the Working Group and the community surrounding it.  A great WGO would do all that and also produce information that helps expand that community.  They could publish quick how-to’s, for example, concentrating on either current or near-future specifications.

If you could, please allow me to illustrate my points with a few things that a CSS WGO might do in the course of their duties.  I’ll call this CSS WGO “Bob” to make the example less clumsy.

Recently, Bob’s been seeing a lot of calls on blogs for an “ancestor” selector.  This would be something that lets you say, “style this element based on its descendants”, such as styling all links that contain an image without having to class them.  (This idea has come up many times in the past, by the way, but has yet to be added to CSS.)  So Bob brings the “ancestor selector” subject to the WG.  The WG says, “Yes, that’s a very good idea, but it runs aground on the following problems.”  Bob would then put all that into his next summary: “The WG is in favor of adding the ancestor selector, but the following problems prevent its inclusion…”  Bob could certainly also communicate the response directly, through mailing lists or blogs, instead of just putting the response in the summary.  The latter is necessary, of course, but doing both is better.

How is this better?  Because the community knows the WG has considered the idea, where the WG stands on the idea, and the reasons why it hasn’t been accepted.  Everyone knows where the sticking points lie, and can make suggestions to overcome them, instead of just guessing as to why the requested feature hasn’t been adopted.  As for the reasons, they could be anything from “that’s demonstrably impossible in an entropic universe” to “not enough implementors have committed to doing it”.  As long as we know what the roadblock is, we can act accordingly.

Furthermore, Bob might accompany a new version of the Advanced Layout module with a quick how-to article that describes how to do a certain common layout, one that’s very hard to do in current CSS, with the stuff in the new module.  This provides a quick, “wow cool!” introduction to the WG’s efforts, which can energize the community and also draw in new people.

I will readily grant that many WGs have what are effectively unofficial WGOs; in a lot of ways, you could argue that I’ve been a WGO for years, as have several other people, through books and articles and forum participation and blogging and so on.  That’s not enough.  There needs to be someone inside the Working Group who is focused on explaining to the world what the WG is doing and who is explaining to the WG what the world is doing, or at least trying to do.

So that’s the first of my three major suggestions for reforming the W3C: an outreach person for every Working Group.

• W3C Change: Outreach was published on Friday, September 15th, 2006.
• It was assigned to the W3C category.
• There have been eighteen replies.

When I posted about the W3C, a few people responded with, “All right, fine, you’re angry with the W3C.  So what’s your alternative, smart guy?”  A fair enough question.

While I applaud the efforts of the WHAT WG and the microformats community, I’m not advocating a complete dismissal of the W3C.  The basic role filled by the W3C, that of being a central meeting place and coordinating body, is an important one.  It’s also potentially damaging.  Think of it like a central file server at work.  As long as the server is fine, your work can continue.  If it goes offline or, worse, its contents get corrupted, you’re in a very bad position.

When I point to the WHAT WG and microformats, I’m not holding them up as saviors or replacements.  I’m simply drawing attention to effects of the basic problem.  Both communities arose because of the nature and (lack of) speed of the W3C and its work.  We could argue about whether or not they should replace the W3C, but the simple fact is that had the W3C been more responsive and in touch with developer needs, they would never have existed in the first place.  They wouldn’t have had to exist.

If the W3C can get back on track, I wouldn’t want to see it replaced.  If it can’t, then it will be replaced, no matter what I or anyone else has to say.  That doesn’t mean it would cease to exist, of course.  It would simply become less and less relevant.  I have some ideas about how the W3C might avoid such a fate, but they aren’t things that I can cover in a single post.  Instead, I’ll do it in three parts, and the three topic areas I’m going to address are:

• Communication between Working Groups and the public
• Working Group membership and chartering rules
• The basic operating model of the W3C

No small potatoes, those.  It will be interesting to find out what people think of my proposals for each.

• W3C Change: Introduction was published on Thursday, September 14th, 2006.
• It was assigned to the W3C category.
• There have been eight replies.

I swear I haven’t forgotten the W3C thing.  Life has just gotten very (and largely unexpectedly) overwhelming of late, and I’ve been falling further and further behind on everything.  To make matters worse, the ideas I want to put forth regarding the W3C are really too long for a single post, no matter how much time I have available.  In fact, I think it’ll take three posts.  I hope to write those soon.  Then again, I’ve been hoping that about a lot of things recently, as my tax attorney and at least two editors can attest.

Before I let it slip any further away, though, I do want to belatedly mention that An Event Apart Austin is open for registration.  Also, this is a great opportunity to mention actual timely news: we’ve just this evening announced that our special guest speaker in Austin will be none other than Molly Holzschlag, who will be giving a talk on designing from the content out.

Don’t miss it, ya’ll!

• Running Toward Austin was published on Tuesday, September 12th, 2006.
• It was assigned to the Speaking and Standards categories.
• There has been one reply.

In my head, at any rate, it was Jeffrey‘s angry post that kicked off the latest round of posts about consortium contretemps, even though Jeffrey’s post was triggered (at least in part) by a message posted to the fairly obscure public-qa-dev mailing list by Björn Höhrmann, detailing his reasons for leaving the W3C.

A little over a week later, there came a semi-rebuttal by Molly over at the Web Standards Project, where she talked about a new spirit of “opening up to new things”, like adding “at least one classically trained artist and graphic designer” to the CSS Working Group (a role that’s been more or less vacant ever since Jeff Veen left the WG over half a decade ago).

That’s great to hear, but what’s perversely fascinating to me is that in that very same post, Molly herself lists the reasons why Jeffrey’s anger is in no way misplaced:

Am I defending the W3C’s slow-to-move process or its over-bureaucratized administration? Its lack of attention and sensitivity to gender (count the women, go ahead, dare you) and racial diversity, its frightening disregard for the real needs of the workaday Web world? Oh no, nor would I want to.

It’s that last point that lends the greatest support to Jeffrey’s argument:  “…frightening disregard for the real needs of the workaday Web world”.

What more really needs to be said?  It’s the most concise indictment possible that the first part of the W3C’s mission statement, the fragment they put right on their home page, “Leading the Web to Its Full Potential…”, has been betrayed.

Believe me, I’d prefer things to be otherwise.  I’m still a strong believer in standards, and for seven years (1997 – 2004) put my time and energy into supporting and advancing them as a member of the CSS Working Group.  When I left, it was because I didn’t have the time and energy to contribute any more, and rather than continue to be a deadwood listing on the group’s roster, I left.  But most of the reason I couldn’t come up with the time and energy was precisely what Molly articulated.  I no longer believed in the W3C’s ability to do what it promised, and what I wanted.

But the worst part?  None of this is new.  Look back two years, when David Baron and Brendan Eich walked away from a W3C Workshop in disgust.  To a large degree, both men walked away from the W3C itself at that point—and if you’ve spurred David Baron to turn his back on the web’s central standards body, then boyo, you’ve got some deeply serious problems.

Let’s be frank: a whole lot of people who believe passionately in the web’s potential and want to see it advance fought for years to make that happen through the W3C, and finally decided they’d had enough.  One by one, I saw some of the best minds of my generation soured by the W3C; one by one, the embittered generals marched forward, determined to make some sort of progress.

Perhaps my eyes have become a touch too jaundiced over the last decade, but I’m not sure I could disagree more with what Molly claims near the end of her post:

Jeffrey is wrong in his current assessment of the W3C.

If only that were so.

If the folks at the WaSP believe the Good Ship Consortium is beginning to change course, then I’m happy for them, really; I’ll be even more happy if they’re right.  But when the ship is moving so slowly and has drifted so far out to sea, how much relevance can a change of heading really have?

• Angry Indeed was published on Monday, August 14th, 2006.
• It was assigned to the Rants, Standards, and W3C categories.
• There have been forty-three replies.

A couple of weeks back, I was hanging out in a New York hotel lobby with Tantek, who was either working on his AEA slides or enhancing the overall usefulness of the web in his spare time; I’m not sure which.  On the far wall, a plasma display ran CNN continually, softly, offering up such choice crawl text as “N. Korea Missile Test Fallout”.  One of the stories running was about alleged plagiarism on the part of Ann Coulter.

We got into a brief discussion over whether such people should be rebutted or ignored.  Tantek took the former position, whereas I took the latter.  My stance is probably a holdover from my long years of Usenet and mailing-list participation, where one of my most iron-clad rules is “Don’t feed the trolls”.  Better they starve for lack of attention, that’s how I see it.  Perhaps this is a defensible strategy in the “real world”, and perhaps not, but I will freely admit that it’s one of my default behaviors.

Thus, my first instinct was to completely ignore John Dvorak’s screed about CSS.  Mr. Dvorak is an admitted troll, and so my default tendency is to simply ignore him.  But “troll” is, in my world, an alternate spelling for “fool”, and as Winston Churchill reminded us, one of the great lessons of life is to know that even fools are sometimes right.

So is Mr. Dvorak right?  Not in what he has to say, no, but there is still something there worth hearing.

It turns out that none of his complaints about CSS are really valid, even when you consider only the ones that have a factual basis.  Sure, he can complain about the cascade being confusing, but that’s like criticizing Windows because of all those stupid windows that open up everywhere and get in the way of the desktop wallpaper.  It’s an inherent feature of the system: either accept it and move on, or reject it and walk away, but don’t waste your time complaining about it.  The best part, of course, is where he blames CSS for inconsistent browser implementations, which is rather like criticizing Microsoft because Windows doesn’t run properly on a computer whose processor isn’t compatible with Intel’s architecture.

But step back and let your eyesight blur a bit, and the shape of a worthwhile point begins to emerge.  The closest Mr. Dvorak gets to expressing it, possibly by accident, is this sentence: “Can someone explain to me exactly what kind of ‘standard’ CSS is, anyway?”

I could do so, of course, as could most of you, but that’s not the issue.  What we’re seeing here is the initial reaction of a CSS newbie, not too different from many others when they first begin to style, and all brought closer to home by the high-profile nature of the newbie.  (Whatever you may think of Mr. Dvorak, he has prominence in the industry.)  CSS is not as hard as some make it out to be, but it isn’t easy as cake, either.

A good part of that problem is the natural expectation that all browsers should act the same.  It’s a strange thing to expect if you’ve been in the field long enough, since browsers have never really been consistent on anything, from HTML error handling to PNG support.  But someone who’s coming in fresh is almost certainly going to expect that if they do things a certain way, the result just works.  Why would one expect anything less?

That’s why the Web Standards Project was founded, of course; and its existence, history, and current efforts put paid to Mr. Dvorak’s assertion that nothing is being done.  As I’ve said, none of his individual points are on target.  What his outburst does is remind us of the problem to which so many have grown numb, and which we still—for all the progress that has been made—face on a daily basis.  Consequently, it reminds us to keep advocating for greater consistency between browsers, to praise the efforts of browser makers in that direction, and to help them correct their course when they move in the wrong direction—and to do so constructively, not destructively.  For while we may gain insights from the rantings of trolls, we should never be so foolish as to adopt their tactics.

• Dvoraked was published on Wednesday, July 26th, 2006.
• It was assigned to the CSS and Standards categories.
• There have been fifty-one replies.

I’ve been largely offline for the last couple of days due to an inexplicable failure of my DSL modem.  I was certain that it was another case of the DSLAM dying on me—it’s happened a few times in the past—and when the Covad techs claimed it had to be a modem failure, I was deeply skeptical.  Score one for the topical experts: they were right, and I was not.

While I waited for the replacement modem that I was sure wouldn’t change anything, I was using dialup.  Man, I never want to do that again.  Talk about sipping the Internet through a cocktail straw.  To make it even worse, I was tethered.  To a phone jack.  There was no wifi infusing the house, letting me work anywhere.  It was like having lost a perceptual sense.  It was wrong and confining and I didn’t like it.  No more of that, thanks.  If the Republicans are so hot to amend the Constitution, how about they be useful for a change and add “the Right to Unfetter’d Bandwidth”?

So.  Nothing much happened CSS-wise while I was gone, did it?  No controversies or anything?  Good.

While I may have been getting my bits by carrier pigeon, the AEA team was able to assemble and post a full schedule for An Event Apart Seattle, which includes a session by Kelly Goto on “Designing for Lifestyle”:

As design migrates from the web to mobile devices, our approach must also shift. Learn how companies are using ethnographic-based research to design smarter interfaces.

I’ve seen Kelly speak in the past, and she’s always funny, smart, and relevant.  I’m really looking forward to hearing what she has to say about ethnography and design.

I’ll be offering updated versions of my highest-rated talks in New York, “Hard-Core CSS” and “One True Layout”, and Jeffrey will be talking about selling standards to difficult clients (especially when the client is a boss) and the importance of writing to good design.  All this and Stan too!  If you’re fixin’ to come see us, the early bird deadline is still a ways off, but don’t wait too long.

• When It Rains… was published on Friday, July 21st, 2006.
• It was assigned to the Speaking and Standards categories.
• There have been seven replies.

Over the past year-plus-a-half, S5 has grown from a small hack of a compact slide show script written by Tantek Çelik into a relatively complex bit of work.  In the beginning, there was simply a way to take a single document and turn it into a series of slides.  I added basic keyboard controls, a navigation menu, and the ability to have the navigation controls show and hide, and then threw it out into the public eye.  People loved it, and with a lot of help from a lot of people, all manner of features were added: slide bookmarks, much better keyboard controls, incremental progress, a notes view, and more.

Despite all this community involvement, though, the code base was in a single set of hands: mine.  Anything that was added to the “official” S5 code was done by me, as time and understanding allowed.  As anyone could have predicted, this has slowed the advancement of S5 over time, and of late it’s brought advancement to a near standstill as I’ve struggled to keep up with other demands.  The only thing I’ve added since 1.2a2 is the ability to blank the screen by hitting the “B” key, and that change has yet to become public.

Of course, the code is explicitly in the public domain, so anyone can add to S5—and many have.  ZohoShow, for example, outputs S5 1.1 code.  I’ve seen S5 used for product tours of medical software and board games.  Jonathon Snook added a “live preview” version of the notes view, which I totally want to see in the primary code base.  David Goodger made a bunch of useful Docutils-compatibility additions that I never managed to fold in.  I also know of four different implementations of remote-control functionality, where one person runs a slide show and changes are reflected in remote copies.  This is a feature perfect for distance learning, corporate netconferences, and other situations.

And all this time, there was still no way to have those enhancements, or any others, “come home” to the source of S5 unless I did it myself.  Until now.

Thanks to Ryan King, we now have S5 Project, which will be the official home of S5.  Besides the blog and mailing list S5-discuss, there will be a wiki, a source code repository, and a bug-and-feature-request tracking system.  If you’re an S5 hacker, or even a frequent user, please do join the mailing list (I know, I know—another one?) or at least subscribe to the S5Project RSS feed to keep track of what’s going on.  I expect the mailing list to become the place for coders to talk about additions they want to make and bugs they’re trying to squash, even after the bug-tracking software gets set up, and it will be a primary source of content for the wiki-to-come.

While it’s been the case that anyone may add to S5 in their own way, for whatever purpose they see fit, now there will truly be community access to what’s always been a community project.  I hope you’ll join us there!

• S5Project.org was published on Wednesday, July 19th, 2006.
• It was assigned to the Projects and S5 categories.
• There have been ten replies.

Jeffrey wrote yesterday about some Flickr problems he’s having, and while he’s found resolution, his post brought to my forebrain some problems I’ve been having with Flickr.  So I’ll record them here.  Wooo!  Flickr pile-on!

Actually, I really only have one problem, but it manifests itself in multiple ways.  The problem is this: any photo with a privacy setting other than “Public” doesn’t ever show up in Flickr RSS feeds.

Here’s why that’s a problem, instead of a good thing:

• If one of my contacts has marked me as a Friend, and they post a photo that’s visible only to Friends & Family, that photo does not appear in my RSS feed of photos from my friends and family.  These same pictures show up if I go to the “Photos from your Contacts” page on the Flickr site.  In the feed, they’re entirely absent.

• If I post a photo that’s visible only to Friends & Family, any comments made on that photo do not appear in my “Comments on your photos and/or sets” feed.  So I don’t know what anyone’s saying about pictures of my wife and child unless I go to the “Recent activity on your photos” page on the Flickr site.

• Bonus related limitation: only comments appear in my recent activity feed; things like added tags and favorite-photo designations don’t show up in the feeds either.  In fact, the feed link on the Flickr site says “Subscribe to recent activity on your photos” but the only activity shown in the feed is comments on public photos.

There may be other, even more subtle hindrances in that vein, but those are the ones that have annoyed me the most.

So why is it that stuff I want to know about—in fact, the stuff that I probably want most to know about—is only available on the actual web site, and not in the RSS feeds?  Flickr knows exactly what it can show me and what it can’t when I visit the site, but when viewed through the lens of RSS, it suddenly forgets what non-public access I’m allowed to have.  To steal a perfectly appropriate line from Jeffrey’s post:

A user experience mistake like this feels quadruply wrong precisely because user experience is what Flickr typically gets so right.

Update: it seems to be a security thing, as a few people have already commented.  I guess I understand the concern, but it’s hard for me to give it a whole lot of credit: if I were that paranoid about people seeing photos I consider truly private, I wouldn’t put them on a central server that anyone can visit in the first place.  Yes, I’ve withheld some photos from being fully public, but that privacy effort is one security breach or late-night coding goof away from total failure.  (Remember when Amazon accidentally showed the real names of reviewers instead of their account names, thus exposing some authors as having slammed books competing with their own?)  So if my personal “recent activity on your pictures” and “photos from your contacts” feeds were based on long randomly generated tokens, and not the discoverable user IDs, that would seem to be private enough—for me, anyway.  Your paranoia may vary.

• Forgetful Flickr was published on Wednesday, June 28th, 2006.
• It was assigned to the Web category.
• There have been thirteen replies.