Subject:
|
Re: Marking articles as 'read'
|
Newsgroups:
|
lugnet.admin.suggestions
|
Date:
|
Sun, 7 Apr 2002 04:12:54 GMT
|
Viewed:
|
2339 times
|
| |
|
|
In lugnet.admin.suggestions, Ross Crawford writes:
> I *did* read the posts around both those refs, and found nothing about the
> data structure (for marking posts as read) already existing.
It's the same data structure (or could be) as the rating data structure,
that's the point.
> Maybe you could
> point me to it?
It's an inference from this:
> In lugnet.admin.general, David Schilling writes:
> > I think a great side benefit of the article rating system is the fact that
> > articles you've rated are marked with '>>'s. This alone makes it
> > worth rating
> > every article you read, as now you don't have to rely on having your browser
> > remember which articles you've read. Especially useful for when you use
> > multiple machines to read messages!
>
> Which clearly means Todd's sw knows articles we've read, if we've rated them.
IOW, there is a data structure of some sort (doesn't matter what, sparse
array, pointers off the articles, linked list associated with the user ID
that has pointers to articles, whatever, it matters not exactly...) that for
each *article* knows who has "rated" it... (or for each *who*, knows which
articles have been rated and what value they were rated as, depending on how
it was implemented, or if you prefer, how you think about it.)
Although it was never stated explicitly, the existenc of this is a very
clear inference, at least to me, and I implied the existence of such a
structure in remarks that Todd didn't contradict. (IOW how could you rate
posts if there wasn't a structure of SOME sort that could hold multiple
ratings for posts associated with who rated them so they could be
redisplayed? Ratings of posts are invariant across machines, try it. The
post refed was in fact rated by me 2 years ago on a machine I no longer have
and it still displays my rating value)
Extending this from the current meaning of having numeric 0-100 values as
the associations (the ratings values, formerly 0-100 by increments of 10 as
possible choices, now only 0 50 100 but still the same data structure
underneath with different tagged meanings on the surface (Todd described the
change somewhere but that may not be an easy post to find) to include also
encoding whether you've seen an article or not is not necessarily hard. (one
could conceive of using negative numbers as flags or whatever, who knows....)
At the time, when it was suggested that the same structure could be used for
that (by implication) Todd did not say "no it can't because..." which
implies that it could. (rather, he said "it could be made to work" without
being specific about what "it" was, but it seemed clear to me he meant "the
use of the same data structure")
And that's all I want to say about it at this point. Server side post/view
memory could be done if it was desired. It's not infeasible and much of the
structure is presumably in place already because of ratings.
It's also not necessarily a good idea or a bad idea to do so. Let's let this
one lie fallow for a while now, OK? I've belabored it more than I meant to
as it is, but you did ask.
|
|
Message is in Reply To:
 | | Re: Marking articles as 'read'
|
| (...) I didn't mention code change size. (...) I *did* read the posts around both those refs, and found nothing about the data structure (for marking posts as read) already existing. Maybe you could point me to it? (...) Well another minus is that (...) (23 years ago, 7-Apr-02, to lugnet.admin.suggestions)
|
8 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
