W3C home > Mailing lists > Public > public-html@w3.org > June 2009

Re: use of data for "decisions" (Was: Issues of @summary and use of data for "decisions")

From: Smylers <Smylers@stripey.com>
Date: Wed, 24 Jun 2009 12:27:21 +0100
To: HTMLWG WG <public-html@w3.org>
Message-ID: <20090624112721.GJ4461@stripey.com>
Shelley Powers writes:

> data queried from a couple of different sources, neither of which is
> all encompassing, shows that the @summary attribute is used
> incorrectly a significant number of times. ...  Yet the same data also
> demonstrates that HTML tables are, themselves, used incorrectly.

Actually since that data only reports on the subset of tables which have
summary attributes, it only demonstrates that tables which have summary
attributes are nearly always used incorrectly; it _could_ be that tables
without summary attributes are more often used appropriately.

But let's assume we have data which shows near-universal misuse of

> However, a decision wasn't made to eliminate HTML tables. ... That's
> the problem with this process, and this practice: it is applied
> inconsistently, and used more to justify personal opinion than as a
> basis for a "best practice".

Not necessarily.  As well as the misuse it's also necessary to consider
the effect of the misuse, and the apparent author intent of the misuse.

* With summary the only possible effect of the misuse is users being
  subjected to the irrelevant attempted summary.  The author's apparent
  intent was to help such users, and it's either had no effect (if it's
  been ignored) or made things worse.

* With layout tables the main affect is of content appearing, to users
  with graphical browsers, next to each other.  That was apparently the
  aim of the author, and for a large number of users it succeeded.  We
  know that it has downsides (not appropriate for users without
  graphical browers; mingling semantics with presentation), but it's by
  no means as big or as unambiguous a failure as summary apparently is.

And it's worth looking at the effect the misuses have on the genuine

* For most users the valid uses of <table> aren't in any way hampered by
  the widespread misuse; a user of a graphical browser who is presented
  with a data table isn't a risk of not realizing that the data is
  tabular because most tables she sees are for layout.  So despite the
  misuses, valid uses still work for at least some users.  (Yes, it's
  bad that this isn't for all users; that's why authors are told not to
  do this.)
* Whereas with summary the misuses so swamped the genuine uses it looked
  like no users could ever benefit from them -- because a user choosing
  to hear a summary would so likely get something irrelevant they'd
  learn not to bother with summaries, thereby missing the few good ones.
  So there's a distinction between misuse which renders the genuine uses
  (and therefore the feature) irrelevant, and misuse which leaves the
  genuine uses working for most users.

  Obviously this view of summary is mistaken: it doesn't take into
  account screen readers' layout table detection.  But given that
  misunderstanding, the conclusion reasonably follows.

So it is possible to logically distinguish different misuses, and
reasonably draw different conclusions just from the data.  (And it's
possible to alter those conclusion by providing additional data which
materially changes the situation.)  It doesn't follow that the data was
used inconsistently, or that decisions were being made on things other
than the data.


Received on Wednesday, 24 June 2009 11:28:01 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:44:49 UTC