[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: what options do we have was Re: POSIX shell\; bash ash pdksh \& /bin/sh



> 	In your (arguably) laudable single minded devotion to
>  technical purity, you are missing the point of HTML browsers and the
>  web: information transfer.

Yet another reason why this comparison to shells is invalid.

> 	When I write HTML, I validate it against a standard DTD, for
>  that ensures maximal dissemincation of my data; however, when I use a
>  browser, I interested in the information the author has to convey,
>  not their expertisae in coding HTML.

It is the job of the browser to guess how the HTML author wants the
page to look.  It is dangerous for a command interpreter to do the
same thing.  If spelling correction were enabled by default in
non-interactive scripts (and therefore without prompting), I would
be horrified.

> 	So, when I read my sister-in-law's page about the family and
>  pictures of the kids, I want to see what she has written, even if she
>  does not know HTML and uses claris front page that produces
>  excruciatingly bad HTML.

If she embedded a sound file on that page for you, consisting of
her specifying a "secret" number, and you were to visit that
page using a browser that didn't, or couldn't, play that audio
for you, you might have no idea it existed.  First and foremost,
this is the fault of the people or software generating the HTML
in a presumtive manner.

> 	If you foist a browser that fails to display that in some
>  misguided "pure HTML" evangelism, that browser is broken, and so is
>  any distribution that makes it a default.

Indeed, it is preferable for the browser to display what it can
rather than to bitch and moan that there's a tag beyond its scope.

> 	Reliability. *RELIABILITY*. Fault tolerance. Pedantic broswers
>  that whimper and fail at the first ill formed HTML are unreliable,
>  and should be discarded in ivory towers.

Perhaps there are cases in which an indication of some kind might be
preferable.  In any case, this is getting farther away from the
shell issue.

> 	This is not the university, people, we are putting together a
>  real world distribution; and while our HTML pages and scripts are
>  held to the highest standards, our interoreters and browsers need to
>  be more fault tolerant.

Browsers, sure.  If I write a broken bash script, I want it to choke,
not second-guess me.


--  
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org


Reply to: