On Fri, 16 Aug 2002, Chuq Von Rospach <chuqui@plaidworks.com> wrote
>On 8/16/02 9:46 PM, "Nick Simicich" <njs@scifi.squawk.com> wrote:
>
>I do. Think about how an iMap client works (or a usenet client) vs. a web
>page. As you navigate the web page, you're doing a lot of loading new pages
>and refreshing. With iMap and usenet, you download the key header
>information and then the client manipulates it.
>
But what is the purpose of an archive? How is it used by the majority of
users?
It seems to me, perhaps I'm just not thinking out of the box, that an
archive is generally used to find and read a small number of articles
meeting particular criteria. It is not used to systematically read all
the messages it contains. Sending all the headers for an archive of
10,000 messages seems over the top if the user only wants to read three
of them. In this situation we want to assess which messages the user is
interested in and send only those. That means server side processing,
and this is something a web interface is reasonably good at.
And NNTP completely useless at. To select all the messages from Joe you
will have to grab all the headers, process them and then request the
appropriate ones from the server. To select all messages mentioning Joe,
you're going to have to grab the lot to search in the bodies.
IMAP does offer significant improvements in this respect and IMHO has a
lot more potential as a vehicle for archives.
What I'm really getting at is that different modes of use will be
benefited by different systems. I tend to the view that as something
that can be searched for that little snippet of information you really
need right now, web archives with a good search facility are far from
bad. I also think they could be improved upon and still be web
interfaces.
>With web, you have a constant stream of relatively slow updates. With iMap
>and NNTP, you have a slower startup time to load the data, but faster
>browsing. And from a user perception view, that slower startup time can be a
>lot less intrusive than constant small delays.
>
Agreed, but... Firstly I question whether browsing is the usual mode of
using an archive. I often find myself at marc for some reason, but I
don't think I've ever browsed a list there - I search for what I'm
after, skim through the results to see if anything catches my eye as
being spot on my question and go there.
Secondly, is user perception the only thing that we should be concerned
about? What about bandwidth? Sure, all the pretties of the web add to
this, but on the plus side the client asks a question, the server sends
the answer. Set this against the server sends all the data and the
client then asks the question of itself.
>But that can be minimized, I think, though careful use of frames. I'm not
>sure I'd go all the way to the javascript browser, IMHO. You're better off
>offering NNTP or iMap and not trying to rebuild those browsers in
>Javascript.
IMAP I can see, NNTP not.
I've been wondering about experimenting with a read only IMAP folder and
IMP to provide a web interface to it. Maybe one day I'll have time to
experiment...
--
Chris Hastie
References:
|
|
