Posts from April 2009

There’s something about the Diggbroglio that has left me scratching my head:  how is it that so many people are up in arms about the DiggBar when they’ve had nothing to say about the framing bars of StumbleUpon, FaceBook, etc. etc.?

Now, please note that I’m not saying the DiggBar, or any other framing bar, is cool and we should all love it.  I’m not.  I absolutely, completely, totally get all the reasons why framing bars are bad for breaking bookmarking and navigating and search engines and copyright and hijacking content and so on.  But that’s precisely why I’m so confused, because we’ve known for years now that framing bars are bad mojo—and yet StumbleUpon, for example, is based on bars.  There is a browser extension/plugin StumbleUpon thingy you can install, but there’s also a web-based framing bar thing (see this link, for example) that they offer, and I bet people use.  You don’t have to be a member to use it: I hit that link in a browser that allows cross-site frame loading and I get the bar and the page it’s framing, and I’ve never been a StumbleUpon member.  The source shows it’s using iframes to make it happen.  So far as I can tell, it’s not really different from the DiggBar.

So why do we have people writing JavaScript and PHP and Ged-knows-what-else that specifically busts out of the DiggBar framing, instead of busting out of all framing?  After all, site framing is universally agreed to be objectionable; even yet-to-be-discovered life forms orbiting distant stars think it’s a bad idea.  So why is one instance of it being targeted while the rest are tolerated?  Why are we all not just using if (top != self) {top.location.replace(self.location.href);} and other-language equivalents?  Yes, I know, some of you do just that, but why isn’t everyone?

Perhaps because I have yet to eradicate a stubborn streak of faith in the rationality of my peers, I assume that there’s some technical difference here that I’m missing and that, once understood, would let me understand the source of the outcry.  So can someone please explain to me, or point at an explanation that states, the technical ways in which the DiggBar is worse enough than already-extant framing bars that it’s triggered this outrage?  Again, nobody has to enumerate the complete list of the DiggBar’s sins; I understand.  A list of any different and more egregious sins would be just fine, though.

Also, if anyone comes up with a way to bust out of the frames while still preserving the bar—that is, correcting the problems framing bars cause while preserving their functionality for the people who want to use them—that would be extra-cool.  After all, people who use those services like the bars.  If we could let them browse the web the way they prefer while fixing bookmark/SEO/etc. problems framing bars can cause, that would be a win all the way around.

Update 14 Apr 09: looks like Porter‘s trying to keep the bar without the framing.

Update 16 Apr 09: in his post about Digg changing the way the DiggBar will work, John Gruber lists (by way of quoting Digg VP John Quinn’s post about it) the two things that made the DiggBar extra-objectionable (at least in his eyes).  Thanks, John!

• Digging in the Mud was published on Tuesday, April 14th, 2009.
• It was assigned to the Questions, Tech, and Web categories.
• There have been twenty-two replies.

In creating the combo-bar charts for the survey report, I stumbled into an Explorer bug that I didn’t remember ever seeing before, and Google didn’t turn up anything that seemed to be related.  This could easily mean that I’m the only person who ever did something this insane and thus found the bug.  It could just as easily mean that my Google-fu has failed.  Either way, I’ll write it up here so it can enter the collective memory.  (And surely someone has already noted that Google is positively Jungian?)

You can see both the problem and two workarounds if you visit this test page using IE6 or IE7.  In brief, the problem occurred when I had a table cell containing a paragraph with an explicit width set.  I did this through the style attribute, though tests show that for this bug, it doesn’t matter whether you use the attribute or assign it via a style sheet.  Around these explicit-width paragraphs were div elements with width: 200%;, for bar-drawing purposes (it’s a little complicated).  Everything was fine in 99% of cases.  But as soon as the header text at the beginning of the row went to two lines of text, the explicit-width paragraphs doubled in width.  What was 80.1% wide would be drawn as though it were 160.2% wide.

My hopefully understandable reaction was to say, “Whuh?”.  I threw a few hasLayout triggers at the offending paragraph (relative position, zoom, etc.) and got nowhere.  In the end I worked around the problem by telling IE6 and IE7 to not wrap text in the row headers of combo charts.  (The bug is not present in IE8.)

I mentioned all this in my announcement post, and the ever-awesome Dan Wilkinson discovered that the problem could be fixed by setting all of the table rows to, say, height: 3em.  Armed with that breakthrough, I experimented a little more and found that I could actually set the offending table row’s height to 2.75em and still have things work as intended.  Below that and the paragraph widths doubled.

Then I lowered the line-height of the headers to 1 and found that I could take the overall row’s height as low as 2.33em before triggering the bug.  And that’s when it dawned: the bug was triggered by the layout height of the header cell’s content being taller than the content of the cell containing the paragraph, and not explicitly declaring styles that would make the data cell as tall as or taller than the height needed to have the header cell contain its content.

Okay, that’s a little dense.  Let me step through the symptoms and two found workarounds to see if that clears it up:

1. The data cell always has a single line of text.  The bug is triggered by having the header cell go to two lines of text, whereas it is not if the header cell has a single line of text.
2. If the row’s height is explicitly set to a value equal to or greater than the header cell’s content, the bug is not triggered.
3. Alternatively, if the height of the data cell is set or forced to be equal to or greater than the height of the header cell, the bug is not triggered.  This can be done with an explicit height or with added top and bottom padding or by adding top and bottom margins to the paragraphs in the cell.

Some side tests I did indicate that the header cell is not needed to trigger the bug.  In other words, the problem could happen even if there are only data cells in the row.  Furthermore, I found that the width-scaling was directly affected by the width of the wrapping div, and thus the widths were doubling only because I had the div‘s width set to 200%.  If I set it to 150% instead of 200%, then the paragraphs only got half again as wide instead of doubling.  That seems to make sense until you see it in action.  When I say the paragraphs got half again as wide, I mean that instead of being 75% as wide as the div, they were 112.5% as wide as the div.  If I set the div‘s width to 200%, then they were 150% the width of the div.

So.  As I say, this bug does not affect IE8, so that’s good, and it can be worked around in IE6 and IE7, which is even better.  The problem would be if you didn’t know how tall your cells might get—in my case, I can be (reasonably) sure that the tallest a cell’s content will ever get is two lines of text, and by setting an explicit line-height on the headers I can know how tall I have to make the rows in order to avoid the bug.

Another day, another workaround!

• Handling an Explicit-Width Bug in Internet Explorer was published on Saturday, April 11th, 2009.
• It was assigned to the Browsers and CSS categories.
• There have been nine replies.

At last—at long, long last!—the results of the A List Apart Survey 2008 are available, along with the anonymized raw data we collected.

There are a great many reasons why it took so long to get this out the door.  A big part is that it’s almost entirely a volunteer effort, which means it happens in our “free time” (and there the word “free” has a couple of meanings).  I say it’s almost entirely a volunteer effort because the detailed analysis is actually done by a pair of professional statisticians, who are paid for their time and expertise.  They did a great job once more, and did it in a reasonable time frame.  It just took us a while to get them the data to analyze, and then a while longer to take their report and findings and process them into report form.

The biggest change this year is that we’re publishing the results as HTML+CSS instead of a PDF.  This greatly increased the challenge, because it was important to me that the data be presented using styled tables, not images.  That’s easy like cake if all you’re doing is putting them up as visual tables, and we certainly do that for some of the figures.  In the other cases, where we have bar charts of varying kinds, things got difficult.  I managed to devise solutions that are 99.9% effective, and I’m both proud of and frustrated by those solutions.  Proud, of course, because I managed to wring three-stack bars out of table markup; frustrated because of the markup I had to construct to make them possible.  I think this report represents more than half my lifetime usage of the style attribute, but unfortunately there’s no way (using just CSS) to say {width: content;}.

So why not use JavaScript to do that, or to just replace the tables with canvas-drawn charts?  I did consider both, but decided that I would push as far as I could with plain HTML+CSS. 

A few implementation notes:

• I used HTML 5 in order to step around some previously unrealized limitations of HTML 4—did you know tfoot has to come before tbody in HTML 4?  I didn’t.  I did not use elements like header and footer due to known problems in Firefox 2 and related browsers, which mangle pages containing those elements.  Instead, I took the same path Jon Tan recommends, and classed divs using those names for later, easier conversion.

• The tables which underlie the charts do not have summary attributes.  If a group of civic-minded individuals would like to write useful summaries, please let me know in the comments and I’ll let you know how best to submit them.  Similarly, I did my very best to make sure all the table headers had accurate scope values, but if I botched any, let me know.

• I’m aware that Opera shows horizontal scrollbars on most chapters of the report.  This is due to its refusal to apply overflow to table boxes, which according to my recent reading of the CSS 2.1 specification is the correct thing to (not) do.  Every other browser I tested does apply overflow to table boxes, though, which I found most useful.  I tried applying overflow: hidden to a few other boxes, and that got rid of Opera’s horizontal scrollbars, but it also cut off actual content in some other browsers.  I chose a cosmetic problem in one browser over loss of content in others.  The best fix I’ve devised is to wrap the tables in divs and apply overflow: hidden to those divs, but I didn’t want to rush the fix and botch it, so it didn’t make it in time for first publication.  I expect to get it in shortly after publication.

• In a like vein, there are a few combo charts where a bar goes shooting off the right side of the chart in IE7.  This appears to be due to some kind of width-doubling problem that’s only invoked on elements with a style attribute when the row header goes to two lines instead of being just one.  Googling for an explanation yielded no joy, and a lengthy series of attempts to hack around the problem came to nothing.  If anyone knows how to counteract that problem other than preventing the header text from going past a single line, I’d love to hear it.  (Update: I’ve implemented the “fix” of preventing line-wrapping in the report, so there aren’t any off-the-page bars right now, but you can see an example of the problem on this test page.)

• Surprisingly, the charts mostly work in IE6.  The exception is some of the triple-stack charts, where data points overlap when the rightmost sub-bars get too small, and also the double-width bars mentioned in the previous point.  I don’t really have a fix for this short of upgrading the browser, but if somebody finds one, I’d be happy to test it out.

On that last point, if there are questions or suggestions surrounding the implementation of the report, we can certainly discuss them here.  With regard to the survey and report itself, though—that is, the questions asked and the results we’re publishing—please direct those thoughts to the comments section of the ALA article announcing the report.  I’m hoping that we’ll have the 2009 survey up within a few months, so comments on what we asked and how we asked it, what we didn’t ask but should have, and that sort of thing could well have a direct impact on the next survey.  But please put those on the ALA site, where more people are likely to see them.

It’s done, it’s out, it’s yours—both the report and the data, about which I’ll soon write a little bit more.  Read the report, or produce your own report using the data.  Just always know that when we publish these reports, we do not mean for them to be the final word.  No, what we always mean is for them to be the first words, a starting point, a place from which to grow.  What comes next is as much up to you as anyone else, and I can’t wait to see what you do.

• Findings of the A List Apart Survey 2008 was published on Tuesday, April 7th, 2009.
• It was assigned to the Culture, Projects, and Web categories.
• There have been thirty-five replies.