Weblog entry #49 for Steve
See all weblog entries made by Steve.
I'm considering disabling comments on articles over a given age.
They get viewed less, so anonymous commentors are likely to have little chance of getting a reply anyway.
It's a big change though, so I'm not sure if I'm thinking it through carefully enough.
Comments on this Entry
Personally if the article is still relevant, I think commenting may still be relevant, and if the email addresses are maintained, people may still get answers.
[ Parent | Reply to this comment ]
[ Send Message | View Steve's Scratchpad | View Weblogs ]
I guess you are right.
I think I'm imagining this is more of a problem because I posted most of the earlier pieces and get mails every day or two asking for clarification, or updates which take a lot of time.
I will leave things as they are for the moment, although it might be worth allowing comments to be closed on a per-article basis with a reason "This package is no longer in the archive" or similar.
[ Parent | Reply to this comment ]
[ Parent | Reply to this comment ]
...Rob
The American Dream isn't an SUV and a house in the suburbs; it's Don't Tread On Me.
[ Parent | Reply to this comment ]
[ Send Message | View Steve's Scratchpad | View Weblogs ]
From memory slashdot has disallowed comments on older articles for at least the past three or four years....
Still I take your point, spam hasn't been a problem here thus far. I've deleted maybe three comments total, and banned a few IPs that seems to have done the job.
I'm aware that it is a problem which is liable to occur in the future, so I've got vague plans to allow comment moderation - but that is a hard problem to solve. Slashdot does it mostly correctly, but there is a fair amount of overhead and it still isn't perfect.
OSNews.com has recently been flirting with moderation and as far as I can see it is an almost completely broken implementation - where non-troll comments are being moderated down by people who merely disagree with the posters opinion, and there is no comeback at all.
I'm interested in the topic generally, but I have no idea just yet how to do it both fairly and in a distributed fashion.
On the plus side here spamming doesn't cause anything to come "to the top". The only obvious problem is that users weblogs might get full of bad content - although weblog owners may delete any comments they wish to, I have no idea how often this occurs in practice but I can't say I've seen evidence of it happening yet.
[ Parent | Reply to this comment ]
I'm aware that it is a problem which is liable to occur in the future, so I've got vague plans to allow comment moderation - but that is a hard problem to solve. Slashdot does it mostly correctly, but there is a fair amount of overhead and it still isn't perfect.Perhaps a bayesian filter such as bogofilter or spamassasin is of some use here?
--------
Felipe Sateler
[ Parent | Reply to this comment ]
[ Send Message | View Steve's Scratchpad | View Weblogs ]
That is an interesting approach, although I'm not too sure how well it would work given the variation in comments.
(e.g. some are short, some are long, some are plain text, some contain commands and output, etc).
I shall go search CPAN to see if there are any existing modules which can be used..
[ Parent | Reply to this comment ]
[ Send Message | View Steve's Scratchpad | View Weblogs ]
Algorithm::NaiveBayes looks reasonable.
I guess we just need to feed it a corpus of "good" + "bad" comments, and allow it to label appropriately. Allowing users to nominate a comment for miscatagorisation is going to be problematic though - as this is an avenue for abuse.
(It is a given that the sytem is liable to be wrong sometimes...)
[ Parent | Reply to this comment ]
Allowing users to nominate a comment for miscatagorisation is going to be problematic though - as this is an avenue for abuse.Of course. I remember an old weblog from you in which you said:
There are people that I've "met" here, and gone on to have interesting email discussions with.Perhaps you could use some help from them (of course if they want to)?
Likewise there are people that I recognise as being both interesting and technically competant. I know when I see their name in a discussion that their contribution will be useful.
--------
Felipe Sateler
[ Parent | Reply to this comment ]
[ Send Message | View Steve's Scratchpad | View Weblogs ]
Yes there are a lot of people "here" that I'd trust to do a good job.
I guess the issue is whether they would both have the time and be interested.
I'd rather setup something "perfect" which all users could use - than rely upon some people having time. That way nobody would feel obligated to do the job.
All this is academic right now though ..
[ Parent | Reply to this comment ]
Instead of prohibiting comments on the grounds of invisibility, I would think about giving them more visibility. The obvious solution that comes to my mind (although pros/cons have to be thought out) is to put a section on the front page that lists the articles that have the newer comments on them, a la 'Recent Weblogs'. So you could have something like 'Recently Commented Articles' which links you to the 5-10 articles that got new articles. Of course, articles should not be repeated (or else the list would be always full of the 2 newer articles). It's not a perfect solution, as it means putting even more stuff on the page, but I think it is a reasonable one.
------
Felipe Sateler
[ Parent | Reply to this comment ]
[ Send Message | View Steve's Scratchpad | View Weblogs ]
That would seem to me to encourage spamming - in the sense that a malicious user will know they can push an article and their newly added comments to the front, where they could be more easily seen.
(As thing stand the site admin(s) can see a special page which shows the most recent 10/20/50 comments posted - and that is used to keep an eye upon things.)
I think the responses thus far have convinced me that it isn't a good idea to deny comments on older articles, so I'll leave it be for the moment and think on it a bit longer.
[ Parent | Reply to this comment ]
Solves Steve's issue of not wanting to be contacted about old articles, although will he remember to do it before the comment is made ;-)
I'm sure it needs careful thought to get it right.
[ Parent | Reply to this comment ]