Author Attribution For Alexandra

page 1 of 212next »

Author Attribution For Alexandra

thedeadlymoose 27 Apr 2015 01:48

Making this thread prompted by Pixeltasim's post in this thread:

I could implement what you suggest, but it might take a while. The .au function is extremely complex at this point do to me doing only an okay job the first time around (mostly because I'd never really attempted anything that complicated with the bot before) and just adding random stuff as time has gone on. If I were to add the requested features I'd probably just rewrite the .au function from scratch. If the implementation can be agreed on by Sunday morning, I could probably just sit down and get it over with, but I would like to know exactly what people want me to add. If people are on the fence about this I'd like to know as well, I'd prefer not to invest a considerable amount of time into this if people don't really care.

EDIT: To review, we want another option to add a user as a translator, and only the original or rewrite author receives the upvotes?

Too late to reach the Sunday morning deadline, but let's discuss what we want from author listings.

Off the bat, some things I'd like:

• Ability to credit things to "Author Unknown". (unless we have that and I didn't know)
• Ability to credit things to multiple people.
• Ability to credit things to "No Author" (mostly for staff and system pages that someone created).

…and a request for someone besides Pixeltasim: I'd like staff to go through and make note of all attributions, once Pixel updates the function. Rewrite probably has dibs on assigning someone here. If not, then maybe someone on Community, or jr staff.

Re: Author Attribution For Alexandra

Pixeltasim 29 Apr 2015 01:44

I would like that note that you can completely override the system pages right now to put No Author or simply Staff if you felt like it. Unknown is the author name I use for deleted or anonymous accounts. When I re-write it I think I'll be able to make it much easier to do collaborations, rewrites, and translations, as the current system was designed for single authors and I sort of just hacked the rewrite and override stuff in there. And yes, having someone keeping track of all the attributions or having a better organized system for that would be great, as right now I still get regular requests to add attributions, which isn't a huge deal but part of the point of having the override page hosted on O5 is so other staff can add to it as well, which people have done a great job doing but it's mostly on a as it's reported or noticed basis.

Re: Author Attribution For Alexandra

Decibelles 27 Apr 2015 07:23

Still stand strong by the option to add translator notes to Alexandra (IE, originally written by [x], translated by [y]). I PM'd Gene R about it and he said he'd get around to making a post on here, and I'm hoping he does soon.

Re: Author Attribution For Alexandra

Dr Reach 27 Apr 2015 09:25

I can commit to this on the SCP-ES' behalf, if the translator notes are included.

Re: Author Attribution For Alexandra

anqxyr 02 May 2015 08:48

So, since we're moving into expanding this thing and using it as a full-blown replacement of the wikidot's absent authorship/ownership system, let's just go all the way with it.

(I think the fact we're doing it is great, btw)

• Instead of current :override: thing, make it a 3-column table: Page, User, and Type. User would just have the username, and Type would have one of the possible override types.
• In addition to the currently recognized 'author' (currently implicit) and 'rewrite_author' types, I propose the following:
  • 'collaborator' for articles with multiple authors. Also note that in this case the 'default' author (the user who created the 0th revision of the page) should be listed as collaborator too. The alternative - implicitly treating the 'default' author as a collaborator if at least one other collaborator is listed on the overwrite page - is in my opinion too fragile and would break under more atypical situations.
  • 'rewrite_collaborator'
  • 'owner' - for cases where the owner isn't the author. E.g. author has left the site and transferred ownership of their articles to another user. If owner is not listed, it falls back to the highest priority user linked to the article (see below).
  • 'translator' obviously.
• We need to formalize the canonical hierarchy of types. In case an article has an author, a rewrite_author, and a translator, who gets the final say about what to do with the article? I think a good approach would be rewrite_author > author > translator, but we can discuss alternatives if people want to.
• Move this page to the wiki proper and allow all users to edit them. If they fuck it up, staff can step in, but in general users should be able to change stuff like this about their own articles without contacting staff.

Pixel, you have anything to say about this? I'm asking you specifically since you're the one who'll be left with the work of implementing it, since Alex is where this would be most immediately visible for normal users. So if you have concerns/suggests, I'd like to hear them. For my part, all my rewrite-dealing code would be trivial to expand to include the new types.

Also, anyone else who has things to say about this, please do.

Re: Author Attribution For Alexandra

Decibelles 10 May 2015 08:11

We need to formalize the canonical hierarchy of types. In case an article has an author, a rewrite_author, and a translator, who gets the final say about what to do with the article? I think a good approach would be rewrite_author > author > translator, but we can discuss alternatives if people want to.

I think staff has to be included in here somewhere, as well as owner. So like this:

owner > rewrite_author > author = translator (in the case of articles being brought over from the other wikis. In this case, if the translator can get in touch with the original author, that's what they should do. Otherwise, their prerogative.) > staff

Move this page to the wiki proper and allow all users to edit them. If they fuck it up, staff can step in, but in general users should be able to change stuff like this about their own articles without contacting staff.

The users would have to know exactly what to do and how, and how it affects Alexandra. Given that it's also open to vandalism or mistakes people won't notice for a while, especially if they're not familiar, I don't know if I could support this. Certainly, I'd like a suggestions feature or option, especially in case someone isn't on IRC or the wiki at the moment (which does happen). Not a full-fledged option on the main site for users to edit and Alex to grab from.

Re: Author Attribution For Alexandra

thedeadlymoose 28 May 2015 22:02

Any further thoughts on this?

Re: Author Attribution For Alexandra

Decibelles 28 May 2015 22:59

Ability to credit things to "Author Unknown"

I have now done this. Horray! (Just under Unknown Author.)

Ability to credit things to "No Author" (mostly for staff and system pages that someone created).

I support this, though I know little about what pages would be needed to credit those as.

Ability to credit things to multiple people.

I also support this, and I know Pixel in chat says he was planning on improving the author function?

I also agree with my previous posts and anqxyr's post. Ideally, would like more feedback on this from other people in case anyone here missed something, but I think everyone in this thread has covered the main topics?

Re: Author Attribution For Alexandra

thedeadlymoose 29 May 2015 08:10

Honestly, I am happy with whatever consensus you and anqxyr have, because you guys have put the most thought into this. Assuming Pixel likes the consensus.

Re: Author Attribution For Alexandra

(account deleted) 29 May 2015 01:44

I still maintain that the "No author" account should be called some variation of "Nobody"

It'd be nice if Nobody itself wasn't already taken but eh

Re: Author Attribution For Alexandra

anqxyr 29 May 2015 08:18

I don't really have anything to add beyond what I said earlier.

I still want to hear Pixel's take on this, since he's the one who'll be doing all the work with Alex. While this page is useful for us on its own, Alex is where its effects are most visible.

Oh, actually, I do have one thing to add. On the matter of moving the page to the main site. Users can currently edit titles of their skip, and this can potentially break Alex, and on few occasions did. I don't see how moving this page on the main site would be in any way worse. If anything, the table structure is more rigid and thus harder to mess up. So I'd still like for the page to be moved.

Re: Author Attribution For Alexandra

anqxyr 03 Apr 2016 13:12

Well then, bumping this again because jarvis. I'm going to concisely reiterate my points, most of which I stated last year, but some of them changed since then, and there are some new ones.

1. Leave the current page as is. Create a new page and transfer the data to it. Delete the old page after 1-2 months.

This is in order not to break anything that currently relies on this page. I can probably fix everything I have in ~10 minutes, but I don't know what else is out there that parses this page (e.g. that X-com-like Foundation game, maybe? Who knows), and I think it would be nice to offer a grace period.

I like "relations-metadata" for the url of the page. It's not bot-specific, and it reflects the fact that it's not about authors anymore, but about all the possible relations between pages and users. Another possibility is "link-metadata", but that may be confused with url links.

2. Make the new page on the main site, not on 05.

I don't see why we can't allow authors to record their own rewrites and such. I don't find the argument of "they will break things" to be sufficient. The new format for the page will basically be just a table, and that's it. When a table is broken, it's really noticeable. Most wiki members can edit a table without breaking it just fine. If anything does break, I'll make jarvis ping the ops about it, and they can go and revert the last edit.

3. New page format.

Page Name	User Name	Relation Type	Date
example_page	Alice	original_author
example_page	Bob	rewrite_author	2012-01-03
example_page	Carol	rewrite_author	2015-09-10
another_page	David	coauthor	2012-10-20
another_page	Emily	coauthor	2012-10-20
translated_page	Фёдор	original_author
translated_page	Bob	translator	2015-03-22
abandoned_page	Alice	maintainer	2015-09-15

This should help illustrate what I have in mind.

example_page was rewritten twice by different people. For the purposes of the !author command, Carol adds the page's rating to her net total, while Alice and Bob do not.

another_page was co-written by David and Emily. Co-written pages cannot have an original_author relation with any user. Both David and Emily add the page's rating to their net totals.

translated_page is pretty self explanatory. Both users add the rating to their totals.

The author of the abandoned_page is no longer active on the site. Since they aren't listed in the table, they default to the user who made the 0th revision. Alice wasn't involved in the creation of the page in any way, and doesn't add the rating. However, before the original author left, they specifically granted the rights to make decisions about the page to Alice. Thus she is listed as a maintainer.

Note that when I'm talking about adding rating to users' net totals, this is just one example of how this metadata can be used in various circumstances. Another example could be to automatically generate and update translated-works hubs if we ever get enough of those that manual updating becomes cumbersome.

The date column is mostly optional, and likely will not be used for anything, at least at first. The one thing I can think of where it can be useful, is to resolve priorities in cases of pages with multiple rewrites (e.g. example_page). However, at the moment I don't even know if we even have a single one of those, so *shrugs*. Back to the date column however. I think it's a nice thing to have, even if it's not very useful, and it doesn't cost us anything. And it's better to have it and not need it, than to need it and not have it.

4. Recognized relation types.

In general, any arbitrary string in the Relation Type column should be parsed and saved. If they don't fall into one of the recognized values, they can then just be ignored.

So, here's some types that I think should be applicable, in a vague order of how much rights they have to the page:

• rewrite_author
• rewrite_coauthor
• original_author
• coauthor
• maintainer
• comaintainer
• translator
• cotranslator

This covers every situation I can think of. If anyone has any other suggestions, I'd be happy to hear them.

---

Okay, I think that's all. I probably forgot something, I'll add it later if so. For now, I'd like to hear what other people think about each point. If everyone is in agreement, I'd like to create and fill out the new page in a couple of days, and add jarvis support soon after.

Last edited on 03 Apr 2016 13:56 by anqxyr Show more

Re: Author Attribution For Alexandra

LurkD 03 Apr 2016 13:20

I really don't have much input on this so it looks good to me to get it rolling. Once it's implemented I'm sure there'll be more to add that we'll figure out later in practice.

meant to reply to anqxyr's post

Last edited on 03 Apr 2016 13:21 by LurkD Show more

Re: Author Attribution For Alexandra

Crayne 03 Apr 2016 16:28

I'll edit this post with my thoughts tomorrow.

Okay, so I forgot and made new post. I claim Monday morning.

Last edited on 04 Apr 2016 07:57 by Crayne Show more

Re: Author Attribution For Alexandra

RJB_R 03 Apr 2016 16:52

This seems like a fine idea, although the fourth point seems like it's a much bigger proposal than the other three points and probably merits further detail and discussion.

---

Whether you like it or not, history is on our side. We will bury you!

Re: Author Attribution For Alexandra

Quikngruvn 03 Apr 2016 22:16

I can behind everything anqxyr proposed, with one minor potential change. I don't see why can't put the attribution page on the main site, but I'd prefer if it were locked to admins and mods. (This is not a deal-breaker, though.)

Re: Author Attribution For Alexandra

Decibelles 04 Apr 2016 00:54

1) I'm fine with this.

(e.g. that X-com-like Foundation game, maybe? Who knows)

If anyone is able to get in touch with Riscvul to confirm as the move happens, it would be nice.

I like "relations-metadata" for the url of the page. It's not bot-specific, and it reflects the fact that it's not about authors anymore, but about all the possible relations between pages and users. Another possibility is "link-metadata", but that may be confused with url links.

I think something with attribution in the url would help to make it clear what it is.

2) Also fine with this.

3) Questions here. A lot of these questions might have self-explanatory answers but I wanna ask them anyway to make sure it's out there and people understand what's happening with them. Some of these questions are also assuming, based on chat/forum posts, that an author details page (.ad) will happen in the future. If this is incorrect, please feel free to tell me so.

For the purposes of the !author command, Carol adds the page's rating to her net total, while Alice and Bob do not.

Will the page show up in someone's .ad page even if it was rewritten and even if it doesn't count to the net total? Will the !au command indicate they have pages rewritten?

another_page was co-written by David and Emily. Co-written pages cannot have an original_author relation with any user. Both David and Emily add the page's rating to their net totals.

Will this functionality support more than two co-authors? If I remember correctly, from what I found, some pages have up to four or even five authors who worked in them (maybe four but I thought I remembered one with five authors). And then you have a couple skips where a rewrite was done by two co-authors (I know Dexanote and Anborough did one like that, off the top of my head, as did Dexanote and Roget). How will this implementation work?

translated_page is pretty self explanatory. Both users add the rating to their totals.

Their .ad page will mention this as well, correct? Will it count towards their avg rating as well if they've done non-translation work? Or if they are the same author as the original? For example, Reach's 179 was originally written by him on the Spanish site.

The author of the abandoned_page is no longer active on the site. Since they aren't listed in the table, they default to the user who made the 0th revision. Alice wasn't involved in the creation of the page in any way, and doesn't add the rating. However, before the original author left, they specifically granted the rights to make decisions about the page to Alice. Thus she is listed as a maintainer.

I don't understand this. So, for example, Salman Corbette has left Roget as the maintainer of nearly all his pages. So:

1. Will Salman still have these pages count towards his net votes, average rating, etc.?
2. Will these pages count for Roget?
3. How will this be indicated in the !au command? Or the !ad command?

Another example could be to automatically generate and update translated-works hubs if we ever get enough of those that manual updating becomes cumbersome.

What we've been working on was this system, where the English list would be automatically updated every certain amount of time, while people translating works and posting them here would be inputting them manually. Do you think you have a way of updating this automatically, especially since we're not using the native number/suffix for when they post their translated works?

4) I don't understand this section at all. Were these points not covered in 3)? If not, can you clarify what this means?

Re: Author Attribution For Alexandra

anqxyr 04 Apr 2016 07:27

1

I like "relations-metadata" for the url of the page. It's not bot-specific, and it reflects the fact that it's not about authors anymore, but about all the possible relations between pages and users. Another possibility is "link-metadata", but that may be confused with url links.

I think something with attribution in the url would help to make it clear what it is.

attribution-metadata is fine too

2

:D

3

For the purposes of the !author command, Carol adds the page's rating to her net total, while Alice and Bob do not.

Will the page show up in someone's .ad page even if it was rewritten and even if it doesn't count to the net total? Will the !au command indicate they have pages rewritten?

*All* pages connected to the user will show up in their !ad, even if they're not involved with them anymore and they're not counted anywhere else at all.

More specifically, the format for pages in the new !ad page will be something like this (for Alice):

Title	Rating	Tags	Link	Created	Relation
Example Page: The Exampling	+28	tale, are-we-cool-yet	example_page	2010-12-31	original_author

!au will not show pages if they're not counted into net rating. If Alice wrote two pages ever, awesome_page and example_page, !au for Alice will show that she has written a single page. Same for Bob. !au for Carol will count example_page.

Will this functionality support more than two co-authors?

Yes, in general you can have any number of anything in there.

where a rewrite was done by two co-authors

See my reply to #4 later in this post.

translated_page is pretty self explanatory. Both users add the rating to their totals.

Their .ad page will mention this as well, correct?

Yes, !ad will include everything.

Will it count towards their avg rating as well if they've done non-translation work?

I expect so, unless people are opposed to it.

Or if they are the same author as the original? For example, Reach's 179 was originally written by him on the Spanish site.

That's not something I expected, but I'll work it out. The sane thing to do would be count each page once per user even if they have multiple relations to it.

The author of the abandoned_page is no longer active on the site. Since they aren't listed in the table, they default to the user who made the 0th revision. Alice wasn't involved in the creation of the page in any way, and doesn't add the rating. However, before the original author left, they specifically granted the rights to make decisions about the page to Alice. Thus she is listed as a maintainer.

I don't understand this. So, for example, Salman Corbette has left Roget as the maintainer of nearly all his pages. So:

1. Will Salman still have these pages count towards his net votes, average rating, etc.?
2. Will these pages count for Roget?
3. How will this be indicated in the !au command? Or the !ad command?

1. No.
2. Yes.
3. It won't be included in !au. It will be included in !ad for both Roget and Salman in a normal way. It will also show up when looking up the summary for this specific page (e.g. via !scp-5678 or !search) as "written by SalmanCorbette, maintained by Roget".

In general, maintained pages will not count for the maintainer in stats or calculations or (in the future) graphs made about these pages. They will count for their next priority user instead (also see #4 reply below). The only things were "maintainer" will show up is in !ad and in page summary lookups.

The "maintainer" position is less for jarvis and statistics and such, and more for general record-keeping and so that people would know where they should go if they want to do something with the page and need someone's permission.

Another example could be to automatically generate and update translated-works hubs if we ever get enough of those that manual updating becomes cumbersome.

What we've been working on was this system, where the English list would be automatically updated every certain amount of time, while people translating works and posting them here would be inputting them manually. Do you think you have a way of updating this automatically, especially since we're not using the native number/suffix for when they post their translated works?

I haven't really looked into it deeply. I mentioned it more as a future theoretical possibility of something the metadata page can be used for. Thinking about it some more now, when the new metadata page will go live, jarvis will know

1. the original author of the page
2. the translator of the page
3. the foreign branch the page came from (via tags)

What he will not know, is

1. the url of the original page.

So if we decide on a standardized way to store this information (bad examples: as a tag; linked at the bottom of the translation; written in a marker on a post-it note somewhere), then yes, autoupdates are definitely possible.

This is, however, is very low priority atm. With the number of translations we have right now, it most likely will be much easier to maintain it by hand. If the volume increases, then we can revisit the issue.

4

4) I don't understand this section at all. Were these points not covered in 3)? If not, can you clarify what this means?

They were partially covered in #3, yes. What this section is for, is three things:

1. To include some possible relations that weren't mentioned in #3 (e.g. rewrite-coauthor, which answers your earlier question about Dexanote/Anborough and Dexanote/Roget rewrites)
2. To establish an order of precedence in which jarvis will recognize the users. For example, if a page has both original_author and rewrite_coauthor, to whom should the rating go? rewrite_coauthor is higher in the list than original_author, so they get the bump to their net total, and the original_author get nothing but a mention in the summary.
   
   Note that as I mentioned in the original post, this is only a vague order. Each particular combination of relations and statistics they're used in will be much more nuanced and will have to be decided on a case by case basis. Nevertheless, I think it's useful to have at least a rough outline like this.
3. So that people could suggest other possible attribution options I didn't think of.

---

Hope that answers your questions. Also I love these wall-of-text posts.

Re: Author Attribution For Alexandra

Decibelles 04 Apr 2016 08:05

1. No.
2. Yes

My issue here is that it wasn't Roget who wrote this stuff (except if he did co-write any of it with Salman but that's anecdotal), it was Salman. So taking those votes and averages away from him and giving it to Roget, even if Salman isn't coming back, feels wrong. I may be misunderstanding you but that's what it appears to be. I think Roget, when searching for Salman's stuff with !au or !ad, could be reasonably brought up as maintainer¹, but again, it wasn't Roget who wrote that stuff. It would also hugely inflate his article count².

In general, maintained pages will not count for the maintainer in stats or calculations or (in the future) graphs made about these pages.

Doesn't this contradict 2. then?

They will count for their next priority user instead

&

Note that as I mentioned in the original post, this is only a vague order. Each particular combination of relations and statistics they're used in will be much more nuanced and will have to be decided on a case by case basis. Nevertheless, I think it's useful to have at least a rough outline like this.

I think before this page goes up on the main-site we really need to hammer this order out. We don't need to account for every situation of course, but I think having something more than a vague outline would be good for both a regular user and staff to know and get used to.

---

More specifically, the format for pages in the new !ad page will be something like this (for Alice):

Suggestions and questions, because now I wanna nitpick your brain here for this kind of thing well in advance.

+ NITPICK FOR THE NITPICK GODDESS

Will supplements be there too? Some supplements are essentially treated as extraneous material (and some are very weird, like this thread sirpudding posted in the proposals subforum). Then you have stuff like minmin's exploration log from the MTF contest where they're essentially tales (as well as the 093 stuff). Will those be counted alongside everything else?

Will essays? I can understand why not, but essays live and die by the vote too, as well as being just as much of an effort as anything else on the site to write. I could see the potential in that. I'm also assuming GoI formats too? Will hubs? I think in the last .ad update, hubs were removed from there, but they have their purpose!

And finally, will we have an .ad page for our favorite author Unknown Author? Or will those not count? :(

Any other nitpicks and specific examples of things to include on the .ad page can be sorted out later!

One last thing; I would like Pixel to chime in on his thoughts as well, given the work he put into making the list originally and any ideas/plans he planned to have before Alex went down.

Footnotes

1. I can imagine the pings in chat could get annoying with it though.

2. As if he needs more of them!

Re: Author Attribution For Alexandra

anqxyr 04 Apr 2016 08:13

1. No.
2. Yes

My issue here is that it wasn't Roget who wrote this stuff (except if he did co-write any of it with Salman but that's anecdotal), it was Salman. So taking those votes and averages away from him and giving it to Roget, even if Salman isn't coming back, feels wrong. I may be misunderstanding you but that's what it appears to be. I think Roget, when searching for Salman's stuff with !au or !ad, could be reasonably brought up as maintainer1, but again, it wasn't Roget who wrote that stuff. It would also hugely inflate his article count2.

Oops, sorry, it should be the other way around. Yeah, as I said later on in the above post, maintainers should receive any ratings or stuff like that.

Re: other stuff

!ad will include literally everything. I don't see any reason why it shouldn't. That doesn't mean that everything in !ad will count for net ratings and such.

I'm open to discussing specific priority issues, but I also think it would be fine to resolve them later on as they come up and implemented, instead of deciding on everything before the page goes live. After all, these decisions wouldn't alter the page itself, only how the information from it will be used.

page 1 of 212next »