Subject:
|
Re: Stick in the mud...
|
Newsgroups:
|
lugnet.admin.general
|
Date:
|
Tue, 13 Jun 2000 23:19:59 GMT
|
Viewed:
|
1639 times
|
| |
|
|
In lugnet.admin.general, Kevin Loch writes:
> What's the difference between someone typing in a url into a browser and
> typing a url into a text editor? Not much, except perhaps the person
> typing it into a text editor should consider proper netiquette.
There's a big difference:
When a person types a URL in wrong by hand, it's usually either a mistake or
because they were being lazy (not to imply that there's anything wrong with
being lazy) or because they weren't aware of the actual correct/canonical URL.
In any case, the wrong URL that they have entered exists only in a dialog box
and (hopefully) never gets cut & pasted from there into any fixed form. If
they visit their browser's history file, or bookmark the returned page,
they'll see the correct URL and not the incorrect URL. In other words, the
incorrect URL does not have a propensity to propagate.
Contrast this with when a person writes a URL in a webpage link -- there, it
exists in a "permanent" form which is accessed repeatedly and propagated to
other pages whenever anyone cuts & pastes the URL from the source file instead
of from the location bar after the page loads.
> A good example of the exceptional /foo/bar scenario is lugnet.com/announce.
> There is an ambiguity between lugnet.com/announce and lugnet.com/announce/.
> The funny thing is lugnet.com/announce exists only point out the fact that
> it exists and the exceptional condition has occurred.
Well, no, actually, the page at /announce doesn't really exist -- not any
more than the page at /flurblzorg exists. However, the page /announce
is a valid page label -- which just happens to be empty. When the server
sees that it's empty, it guesses that the user really meant /announce/
since it sees that /announce/ exists. But it can't know for sure because
/announce is a valid page label and because older versions of pages may
exist in the version history for a given page.
> Furthermore, the only relevant content on lugnet.com/announce is a link to
> lugnet.com/announce/ which the user would have normally been redirected to
> if the exceptional page diddn't exist.
But the exceptional page doesn't exist. (It even says that on it.) To be
more precise, there is no page /announce in the system (internally, that
is to say, there is no file /announce.ftx in the document system). The
important point is that there could be, just as there could be a page
/flurblzorg if one were created.
> Seems like a waste of time and resources to me.
>
> Interestingly, lugnet.com returns the same content as lugnet.com/
> as one would expect. Why not create an exception condition for lugnet.com
> also? I bet lots of people link to and type lugnet.com instead of
> lugnet.com/. Shouldn't we set them straight?
Well, for one thing, the root directory is special because there is no higher-
level directory. And furthermore, there's no ambiguity between
http://www.foo.bar/index.zot
and
http://www.foo.bar.zot
as there is between
http://www.foo.bar/blurfl/index.zot
and
http://www.foo.bar/blurfl.zot
because
http://www.foo.foo.zot
isn't valid.
> Of course not. The whole thing is silly. The only time I have ever seen
> an ambiguous condition between directories and files is the ones on lugnet
> that just point to the otherwise unambiguous content.
It's not silly at all. If someone makes a link to some page
http://www.lugnet.com/foo/~1234/bar
which works fine for several weeks, but then the owner of the page later
decides that the page needs its own whole directory and moves the content to
http://www.lugnet.com/foo/~1234/bar/
and sub-pages, it's not clear whether the linker intended to have a link to
the content of the old document (which he/she knew about) or to the content
of the new document (which he/she presumably doesn't yet know about) or to
the content of whichever the linkee prefers, old or new.
> The only concequence is a 301 redirect and an entry in the log file.
> That's a small price to pay for delivering the content the user requested.
Actually, it's not the only consequence. When a server happily allows people
to write both
http://www.foo.bar/blurfl
and
http://www.foo.bar/blurfl/
it is doing people a *dis*service. Sure, it's nice to the lazy people who
don't want to write URLs the right way, or who want to type them in quickly
the wrong way, but it's terrible for people who have to browse webpages with
the wrong links and actually expect link colorings (blue/purple) to work as
they're supposed to. How many web browsers do you know of that color a link
http://www.foo.bar/blurfl
purple (e.g., VLINK) if you've already visited
http://www.foo.bar/blurfl/
but haven't yet visited
http://www.foo.bar/blurfl
?
This whole 301 automatic forwarding business is Really A Bad Thing. It's too
bad that it ever got started. :-/
--Todd
|
|
Message is in Reply To:
 | | Re: Stick in the mud...
|
| (...) /foo/bar/ (...) someone (...) What's the difference between someone typing in a url into a browser and typing a url into a text editor? Not much, except perhaps the person typing it into a text editor should consider proper netiquette. A good (...) (24 years ago, 13-Jun-00, to lugnet.admin.general)
|
85 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
