Comment indexing

[sara]

Title:
Comment indexing

Area:
commenting

Summary:
Build a thingy that will pull the subject lines and direct links from top-level comments on a post and output them in some kind of useful format (maybe a table or a list?) which can be inserted into the body of the post (hopefully under a cut-tag). Bonus points if it can be set to periodically run an automatic update.

Description:
This suggestion is motivated by my total lack of enthusiasm for spending an evening manually indexing a meme. I don't want to do that, that's a database query. Someone should build a thingy so the server can do this job for me.

Specific thingy-implementation-related issues should be conceptualized by someone with more thingy design and implementation experience than I have.

Poll #8881 Comment indexing

Open to: Registered Users, detailed results viewable to: All, participants: 75

View Respondents

This suggestion:

View Answers

Should be implemented as-is.
47 (62.7%)

Should be implemented with changes. (please comment)
1 (1.3%)

Shouldn't be implemented.
3 (4.0%)

(I have no opinion)
24 (32.0%)

(Other: please comment)
0 (0.0%)

Comments

[jumpuphigh]

I love this idea and approve of how the suggestion was written.

Build a thingy...

[denise]

"Thingy" is a very useful metasyntactic variable.

[jumpuphigh]

I agree!

[sara]

...it's a lot less vulgar than the other metasyntactic variable I could have used. *GRIN*

[jumpuphigh]

*snicker*

[sara]

Indeed, this post is substantially less vulgar than the way in which I initially posed this suggestion. *cough*

[azurelunatic]

I just bet! :D

[montuos]

+1 ;D

Comment indexing: sketchy speccing

[azurelunatic]

Automated version:

Retrieves links to top-level comments, and also the subject lines of those comments.
Formats the links, titles, and possibly also timestamp and username of the starter, possibly even current number of replies? timestamp of last reply? idk.
Posts, in a harmonious format (I have a bunch of thoughts about this, many relating to my thoughts on tweet-importers); possibly even by comparing new stuff since last update and highlighting the first $SOCIALLY_ACCEPTABLE_NUMBER of these outside a cut.


"Manual" version, potentially

Retrieves links to top-level comments only. Also retrieves pertinent info mentioned in the automatic version.

Taking the links from step 1, walks down the trees and retrieves all subject lines from these trees.

Presents all the sub-subject lines in some pleasing format for human review (because sometimes you want to use the fic title rather than the prompt title)

Stuffs it into the draft folder, labeled and sorted.

Re: Comment indexing: sketchy speccing

[sara]

Thank you (and yes, I entirely agree about the need for a $SOCIALLY_ACCEPTABLE_NUMBER variable....)

Re: Comment indexing: sketchy speccing

[azurelunatic]

My twitter-importer displays the first 5 tweets; LJ's displays 10; I can't remember offhand the number that the bloke who wrote my twitter-importer's version displays, but it's substantial but not all.

Re: Comment indexing: sketchy speccing

[sara]

I have been known to drop people for crossposting excessively-long uncut Twitter lists. If I wanted to read someone's Twitter output, I'd be...reading their Twitter output. *shrug*

[green_knight]

What a great idea. This is not something I would ever have thought about - I don't get a fraction enough comments for this to be useful - but I can see it being fantastic for busy communities.

Bonus premium feature: enable it so that all readers of a journal/community can view this, rather than it being something that the owner needs to trigger manually and post somewhere. (I can see an 'enable thingy' checkbox here.)

[susanreads]

Wouldn't that be a lot of server load? I think it would make a good paid feature.

[amadi]

Except where it would be most used are exactly the places least likely to be paid: specifically various fannish meme (ahem) communities.

[sara]

*grin* Believe me, the meme that motivated this is on a paid account.

That said, I have no idea what the paid-to-unpaid ratio is on meme comms. Though if you were looking for something to convince meme comms to chip in, well...something that automated what's now taking hours and hours of mod time might just be it.

[amadi]

This is true. Though making indexing a paid feature could also stifle creation of meme comms by those who can't afford a paid account. Interesting issue to balance.

[sara]

*shrug* There'd be nothing to stop people from doing it the way it's being done now, so I don't know how discouraging it would be.

That said, sorting out the business case is not any more my problem than sorting out how to spec the thingy.

[boundbooks]

All things considered, I wonder if looking at things from the point of view of the comm is the wrong way to go about it. Considering the kind of users who participate in kink-memes, is the presence of better/easier memes on DW an incitement to have *them* pay for accounts? I don't think the question really is whether the comms should have to be paid to do this (although they could be), but the bigger question might be whether this makes DW more appealing to people who are potential paid users on their personal journals. :)

[matgb]

Theoretically you could build an S2 layer that'd do this, and paid accounts can view it on any post, etc.

I say theoretically because I don't know I'd know where to start.

[arethinn]

Neat idea (although I'm not sure I'm visualizing the correct kind of post where this would be wanted), although sounds server-loady, definitely paid feature or even premium paid... (Would it be editing and re-editing the top post, thus generating crossposts where applicable?)

[sara]

Like I say, my tendency is to think that the particulars of implementation and paid-vs.-unpaid are issues for The Management, not myself.

[arethinn]

Oh, I know, just musing aloud.

[azurelunatic]

Consider the Oh My God We Need Some Porn In Here Stat Porn Meme (contains textual porn) -- someone reading that might appreciate an index to what prompts existed without going through pages and pages, and of what existed, what had been filled (especially if not a writer and just reading through).

[sara]

*cough* To pick a TOTALLY RANDOM example, yes.

[azurelunatic]

It was the one that sprang to mind, yes. :D

Very loosely paraphrased.

[sara]

"Someone should index your porn meme, Dee. How'm'I supposed to write porn this weekend if I can't find the fucking prompts?"

"You could do it!"

"Oh, fuck that, I'm not manually indexing a porn meme on a Friday night. You own the service, you should build some shit that'll index the porn meme FOR me."

"That's not a bad idea."

"I know, I'll put it in your suggestions comm. Because you totally don't have enough spam in there already."

"You should, actually!"

"Really?!?"

"Yep."

"...righty-o, then."

Re: Very loosely paraphrased.

[azurelunatic]

... ahahaahaha! :D

Re: Very loosely paraphrased.

[sara]

Bog-laziness and dirty-mindedness: motivating technological innovation since the dawn of time.

Re: Very loosely paraphrased.

[azurelunatic]

The story on why there is the link-to-me module on the bottom of the profile page: I was thinking about a problem with OpenID accounts (on the one hand: the internal "ext_#" username needed to be exposed less; on the other hand, people still needed to be able to know the "ext_#" username for OpenID users to link to them and such), and I got fed up of doing that, and fucked off to YouTube to procrastinate. Where YouTube's lovely copy-this-code-to-embed-or-link jumped out at me, and I ran screeching back to Bugzilla with the problem answered and the video forgotten.

Re: Very loosely paraphrased.

[sara]

...you know, we've had that for so long that I sort of forgot we ever didn't have that. *goes and looks* Heyyy, I see what you did there.

[arethinn]

OK, not that far off then. (I was picturing an Anonymous Love Meme, where the index would be by username-as-thread, I guess.)

[azurelunatic]

Oh, those are excellent candidates for indexing as well!

[sara]

*nods* Really, any activity where the action happens primarily in comments would benefit.

[yvi]

Doesn't sound too server-loady to me from my (limited) experience - not more load than people clicking through all comment pages to find something they like, which would also be several database requests instead of one.