- From: Stephen Hay <haymail@gmail.com>
- Date: Wed, 21 Oct 2009 19:49:03 +0200
- To: Giovanni Campagna <scampa.giovanni@gmail.com>
- Cc: news@terrainformatica.com, "Tab Atkins Jr." <jackalmage@gmail.com>, Www-style <www-style@w3.org>
On Wed, Oct 21, 2009 at 6:23 PM, Giovanni Campagna <scampa.giovanni@gmail.com> wrote: > 2009/10/21 Andrew Fedoniouk <news@terrainformatica.com>: >> Giovanni Campagna wrote: >>> >>> 2009/10/20 Andrew Fedoniouk <news@terrainformatica.com>: >>>> >>>> Giovanni Campagna wrote: >>>>> >>>>> 2009/10/20 Andrew Fedoniouk <news@terrainformatica.com>: >>>>>> >>>>>> Tab Atkins Jr. wrote: >>>>>>> >>>>>>> On Mon, Oct 19, 2009 at 11:43 AM, Andrew Fedoniouk >>>>>>> <news@terrainformatica.com> wrote: >>>>>>>> >>>>>>>> CSS tables cannot reproduce <table> layout. >> >>>> >>>> It seems like in mentioned engines table-cell value of display >>>> attribute is changing the meaning of percent length units. >>> >>> Yes >>> >>>> Is such behavior defined somewhere? >>> >>> Informatively in CSS2 17.5.2.2 and normatively in >>> css3-tables-algorithm, available at >>> <http://dev.w3.org/csswg/css3-tables-algorithm>. >>> >>>> So instead of 30% of parent width they use some percents as >>>> sort of flexes. >>>> >>>> How then to set width of display:table-cell element to >>>> 30% width of its container? >>> >>> table-layout: fixed >> >> So we allowed to have sort-of-flexes but without percentages or >> percentages but without flexes. Really the whole design around tables in CSS >> looks like last minute hack. > > If 1998 (publication of CSS2) is last minute, I agree with you, it is > not intuitive in some cases. But I'm not sure we can change it, given > the amount of currently published content for which it works. > >>> >>>> These display:table-*** dances looks like dirty hack here. >>> >>> But they work. >> >> >> Consider these samples: >> >> http://terrainformatica.com/w3/tables-css.htm >> >> I do not think that all this can be considered as something >> ready for prime time. > > Well, that looks really weird. > Maybe Tables 3 or Tables 4 may include a "flexible" table algorithm > which could make more sense. > >> Really I have designed html table engine by myself and >> have implemented Templates in CSS but I failed to get >> idea of half of samples in the document above (questions are in the >> document). > > That's the point of having a normative algorithm for implementers. If > needed, we can add a new one for authors. > >> How all this in principle is solving problems of <table> >> based page designs? > > I don't actually know, I stopped using <table> a while ago (until IE8 > came out with almost full CSS2 support). You should ask the designers > of <table>-based websites. Please don't. The whole idea is to use CSS for layout and to have layout not be intermingled with the content. How it's done on the implementation side, we don't really care. But we need better tools than we have now, and the *author* side of things needs to be something in the direction of Template Layout, because that's the way designers think and work. Designers don't *like* table-based layout in any form. Writing it is a pain. Debugging it is a pain. Changing content is a pain. They like the *control* of it, but that's the extent of it. Tables are a maintainability nightmare. If you (not *you*, but in the general W3C sense) give us the control we need, separate from the content with a syntax designers can relate to, you've changed the face of design on the Web. With layout being one of the biggest problems for designers, I fail to understand the relative low priority. Are shadows and backgrounds simply easier to implement? We need grid-based layout sooner than later. How do we go about getting it? /Stephen
Received on Wednesday, 21 October 2009 17:49:38 UTC