Subject:
|
Re: Specific problem: finding members - general conclusion: principle of least surprise needed on Lugnet
|
Newsgroups:
|
lugnet.admin.general
|
Date:
|
Fri, 14 Jan 2000 14:01:02 GMT
|
Viewed:
|
226 times
|
| |
|
|
Allan Bedford wrote:
>
> Todd Lehman <lehman@javanet.com> wrote in article
> <Fo9Jz6.1q4@lugnet.com>...
> > In lugnet.admin.general, Allan Bedford writes:
> > > As a person who has spent way too many hours on the web over the last • few
> > > years I'm not sure why you claim that this link *ought* to work. I • almost
> > > never find myself trying to guess the directory/URL structure of • someone
> > > else's website.
> >
> > I think Larry is absolutely right. It -ought- to work like that, in case
> > someone happens to type that in if they can't find the page easily (say,
> > they knew it was there and they'd been to it before, but they couldn't
> > remember the URL off-hand).
>
> Todd, my only question about this comment is:
>
> Are you then responsible to your visitors for always maintaining the same
> URL structure for ever and ever? What if... for whatever reason... you
> decide to make some changes. It's your domain name, it's your server, it's
> all yours. So long as people can always find http://www.lugnet.com, isn't
> that enough?
Major sites should avoid re-structuring as much as possible, and should
provide redirects when the structure changes. I am slowly building a set
of LEGO reference pages. Because there is no guarantee that site
structure won't change, I often link to the top page, but then I
sometimes need to give directions as to how to find the item I'm
referencing.
Ultimately part of the problem is that the web is grossly violating good
database design practice. The primary key to documents (and sites) has
meaning. It would actually be best if every document on a web page just
had a serial number, and links would point to that serial number. The
web designer would then be free to put the document in any directory
which made sense to him at the time. Of course there is still the
problem if a document becomes obsoleted. In that case, a simple document
which either re-directs the user to the top level site, or re-directs
the user to the replacement document, or re-directs the user to a deeper
page which might be related (for example, perhaps you had a page
documenting a castle construction. After several years, you decide that
that castle is so poor compared to the current state of the art that you
no longer want to publish it. You might then make the re-direct point to
your general castle page, where people may then browse the castles you
are still publishing).
Of course internet addresses are also a problem. There there are sort of
two primary keys. In one way, the IP address is the primary key, on the
other hand, IP addresses change, so the DNS name is used as the primary
key, but then those change also. Both of these keys have meaning (the IP
address of course tells the routers how to actually find the particular
site, the DNS name places the site into a named hierarchy).
--
Frank Filz
-----------------------------
Work: mailto:ffilz@us.ibm.com (business only please)
Home: mailto:ffilz@mindspring.com
|
|
Message is in Reply To:
16 Messages in This Thread:
 
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
