Gentoo Archives: gentoo-pms

From: Ulrich Mueller <ulm@g.o>
To: gentoo-pms@l.g.o
Subject: Re: [gentoo-pms] The "Feature Availability by EAPI" table is ugly
Date: Sun, 06 May 2012 20:19:14
In Reply to: [gentoo-pms] The "Feature Availability by EAPI" table is uglier than Portage code by Ciaran McCreesh
[Subject slightly adjusted.]

>>>>> On Sun, 6 May 2012, Ciaran McCreesh wrote:
> The "Feature Availability by EAPI" table looks nasty. Once we add > EAPI 5 in there, it won't properly fit on the page.
I never liked that the table is rotated, in the first place. While we're at it, can we get rid of that, too? For example, we could use the dpfloat package and place the table on a double page.
> Here are some ideas:
> Left align the 0, 1, 2, 3, 4, 5 column headers.
> For "default_ phase functions" and "econf arguments", there's too > much column text. We could either replace this with "see text", or > put a quick summary into a different table and say "see table D.2".
+1 (for either solution)
> We could nuke the first column.
The "feature" column? That would leave us only with the "reference" column which is rather cryptic. Also several features share the same reference, e.g., "! blockers" and "!! blockers" are both mapped to bang-strength.
> We could put in some "light" lines on the table. Then for features that > are shared across EAPIs, we can do this:
> | 0 1 | 2 3 | 4 5 | > +----------------------+------------+--------------+-------------+ > |Use Dependencies | No | 2-style | 4-style | > +----------------------+----+-------+--------------+-------------+ > |src_compile | 0 | 1 | 2 |
> possibly with multi-column text centred.
That doesn't look like it would improve the clarity of the table. And you'd have to add lines to separate the columns, which will even take away some of the available space.
> Or we could do it using some kind of background shading on cells > instead?
I fear that whatever of the above we do, we will be discussing this table again for EAPI 6 or 7. It doesn't scale for an arbitrary number of EAPIs. Ulrich