# [00:06] <Bert> It doesn't answer the question I asked on IRC during the telcon: does 'border-radius: 40%' create a quarter circle or is it the same as 'border-radius: 40% / 40%'? :-)
# [00:07] <TabAtkins> Ah, you mean "If there is no slash, then the values set both radii equally." is ambiguous?
# [00:07] <Bert> Minor: the word "length" should be removed from a few occurrences in the the first para. It is not necerrily a length any more.
# [00:08] <Bert> Yes, equal percentage? or equal value?
# [00:09] <Bert> Also: computed value is probably "absolute length or specified percentage"
# [00:10] <Bert> Hmm, not literally that, but something like it, in plural.
# [00:12] <Bert> I think I prefer to make a single percentage make a quarter circle (and the percentage is a percentage of the width). You can always add a second percentage to refer to the height, but you cannot otherwise get a quarter circle.
# [00:18] <Bert> What's more confusing: that 'border-radius: 1em' makes a circular arc and 'border-radius: 5%' does not? or that 'border-radius: 1em' is the same as '1em/1em', and '5%' is not the same as '5%/5%'?
# [00:19] <Bert> I think it makes little difference, but the latter is more powerful.
# [00:20] <Bert> (On the other hand, why anybody would use percentages, apart from 50%, I can't imagine. It makes the corner unpredictable and different for each box and/or each window size.)
# [00:20] <TabAtkins> The latter is more confusing to me. The former seems sensical; all the border-radius components are 1em, or 5%.
# [00:21] <TabAtkins> You thinking what I'm thinking, Bert?
# [00:21] <Bert> Non-circular corners are very rare.
# [00:22] <Bert> Eh, not sure. What are you thinking?
# [00:22] <TabAtkins> % always refers to width, but border-radius can take a keyword value instead of its list of numbers, with only one current keyword - 'oval'.
# [00:23] <Bert> I had that thought, yes. It's simple. But are we sure that's really the only case Håkon needs?
# [00:23] <TabAtkins> I'm not sure. I know it's by far the most *common* case in what Hakon needs.
# [00:23] <TabAtkins> We'd have to ask him if that's sufficient.
# [00:24] <TabAtkins> Of course, we also have to make sure it works for border-image.
# [00:26] <Bert> You mean: Brad may have a different case in mind where he needs percentages?
# [00:27] <TabAtkins> Yeah, and I agree with him that the two cases should work the same.
# [00:27] <TabAtkins> So lock Hakon and Brad in a room and let them hash it out.
# [00:29] <Bert> Well, in that case it's easier to just keep the percentages. :-)
# [00:29] <Bert> (After we remove the ambiguity of a single percentage.)
# [00:29] <TabAtkins> Bah. Surely international travel costs can be justified! This is SCIENCE we're talking about here!
# [01:02] <Bert> Yes, I mean that the x radius stays constant while the y radius varies with the height of a box. (Because boxes vary much more often in height than in width.)
# [01:02] <fantasai> I can imagine wanting percentages against the height for things like buttons
# [01:02] <fantasai> that tend to vary more by width
# [01:03] <TabAtkins> Buttons tend to be squat and vary in width, as opposed to the average CSS box.
# [01:03] <fantasai> In that case you'd put a percentage agains the height and a fixed length against the width
# [01:03] <fantasai> I think we'll see that a lot for tabs, actually, now that I think about it
# [01:05] <Bert> But other than what may be common or not, my main argument is that taking a single percentage to mean equal radiuses gives more funcionality. You can always write two perentages, but you cannot otherwise get circular arcs.
# [01:05] <TabAtkins> I agree that it offers more functionality; that can't be argued. I'm arguing that that's confusing, and think that the extra power granted by it isn't enough to offset this confusion.
# [01:06] <TabAtkins> (And, just for clarification, you can't get circular arcs *that are based on a percentage other than 50% of min(width,height)*.)
# [01:06] <TabAtkins> (As the giant-length hack can give you the 50% case.)
# [01:07] <Bert> My first thought was that it was confusing that a single value did *not* give me equal radii on all sides. By now I don't know what's confusing anymore. It's all equally non-memorable. :-(
# [01:07] <TabAtkins> Yeah, possibly. Push to %h and %w units, and we'll be dandy.
# [01:09] <Bert> Why do you say that you can't get circular except with 50%? In my interpretation of single percentages, *every* single percentage gives circular arcs.
# [01:09] <TabAtkins> (Yeah, %h and %w are a horrible idea anyway, as they'd make circular dependencies ridiculously easy.)
# [01:09] <fantasai> you have to pick something to reference