- From: Rafał Pietrak <rafal@ztk-rp.eu>
- Date: Wed, 18 Jun 2014 09:44:23 +0200
- To: www-style list <www-style@w3.org>
as per our missunderstanding of "off the list" intentions: It was not my intention to talk off the list. funny total missunderstanding :) Then: I'm sending this as a "replay" as I'm not quite sure if the list will retain thread if mail for forwarded instead. so my replay below is actually quoted. W dniu 17.06.2014 23:08, Rafał Pietrak pisze: > Sory for the delay - I was offline for some time; yet, I'd like to dig > this bit little further. > > (I've notice you've taken this reply off the list, so let's continue > this way). > > W dniu 13.06.2014 18:21, Tab Atkins Jr. pisze: >> On Fri, Jun 13, 2014 at 2:28 AM, Rafał Pietrak <rafal@ztk-rp.eu> wrote: >>> W dniu 12.06.2014 22:43, Tab Atkins Jr. pisze: > [-------------------] >>>> to import the position of another element and using it directly loses >>>> this. >>> I'm not exactly sure if I get it. >> Sorry, this wasn't "your" use-case, it was Simon's, at the beginning >> of the thread. > > I sort of assumed that. But still, wasn't able to figure out the case > anyway. Yet, your comments below do explain it sufficiently. > > So, to the point: > > [-----------------] >> I mean what the thread-starter wanted, which is a context menu - if >> possible, they put their top-left corner where you clicked, expanding >> to the right and down, but if there's not enough room, they'll expand >> up or to the left instead. Tooltips do something similar. >> >> Note, though, that neither of these are addressed by "position >> something near this element", because both of those cases actually >> position near the *mouse cursor* at whatever time they started showing >> (upon right-clicking for context menus, upon a delayed hover for >> tooltips). >> >>> Yet, in either case, I cannot figure out what exactly is lost here. >>> In both >>> cases dereferrence of other (fixed, or currently hovered above) element >>> position attribute is the thing that will position the "said tooltip". >>> right? >> Your suggestion would not allow tooltips or context menus to choose >> which direction to expand in based on whether or not they'd overflow >> the screen. >> > > Yes. > > But I don't think it should. > > This use case is "positioning relative to position where right-click > happened", but not in ordynary case, only in case when it wouldn't fit > inside current viewport clipping if rendered "here" ... and have to be > repositioned. This isn't anything like currently available > floats-flexes-etc "repositioning" in css. So, I was wrong: this isn't > just the case of "take a position of another element". > > Now I think, that this use case which could be spoken out as: "here is > a list of other elements, avid covering them in such order/preference". > > In consequence, my conclusion here would be: > 1. position referring to other element position is usefull and would > be welcome, but only usefull in case when repositioning is not necesary. > 2. if referring to position of "other element" is possible, then > referring to other attributes of other elements should be possible, > too. and will be wellcomed even more. > 3. I may be wrong here, but I don't see an "avoid coverage" attribute > in CSS, yet the context-menu "use case" would benefit from such > tremendously. And from the eyars of CSS availability to web-designers > community, I'd say, that this could be yet another enabler for > currently unexpected usage. > > One thing is to "display:block" a context menu, the other is to > "reposition it", when it doesn't fit. IMHO those two should be made in > CSS explicitly distinct, and the better if the later is only hinted to > the brawser (as opposed to "pixel-perfect" specifying by page designer). > > -R
Received on Wednesday, 18 June 2014 07:45:12 UTC