Wikisource:Scriptorium/Archives/2008-06

Proposals

FlaggedRevisions

According to a recent post on the WS mailing list, FlaggedRevisions is now available for any wiki which wants it. This has been a long-anticipated feature. I propose we have this installed on WS.—Zhaladshar ^(Talk) 20:31, 4 June 2008 (UTC)

This is fantastic, and we have waited a long time! Have people here given any thought to which setting and options are best for Wikisource?

In terms of rating pages, my initial thoughts were: "Completeness" | "Accuracy" | "Formatting" each rated from 0-5 with 0 being lacking nearly all text ("Completeness"), not proofread at all ("Accuracy"), or not formatted at all in terms of local standards and aesthetic considerations ("Formatted"). These are just some initial suggestions. It would be great if discussion here could create a basic suggested implementation for Wikisource. Dovi 00:52, 5 June 2008 (UTC)

This was discussed on IRC, and I came away with the impression that FlaggedRevisions is incompatible with Revision patrolling. If that is the case, I think we should be cautious of replacing a working technology (with documentation and tools) with bleeding edge technology that may not provide replacement functionality. John Vandenberg ^(chat) 01:01, 5 June 2008 (UTC)

I think this is too fundamental to leave for IRC discussions that not all are privy to. Although whatever was learned through those discussions should be explained here. Do the current patrolling tools offer anything close to the reassurance value to users that, when available, the page they are viewing is always the one that has been checked for vandalism and rated for quality? Dovi 01:05, 5 June 2008 (UTC)

I am raising the question/concern here because I dont know or recall all of the details discussed on IRC. I am less concerned with what "immediate" impression readers have of enWS, as I hope that they will gain an appreciation for our quality as they browse around. I am much more concerned with our ability to catch the vandalism or degradation of quality as it happens, and yes, the tools in place do a fantastic job of ensuring that unpatrolled edits are clearly visible and able to be addressed as they appear. Also, the unpatrolled edit queue gives us an opportunity to quickly spot newcomers who need a bit of guidance.

I can see that FlaggedRevs is a very useful tool, and would recommend it on any small Wikisource where the number of daily influx of strange edits is lower that on enWS. I would support FlaggedRevs on enWS if there is a similar way to construct a similar queue of edits that need to be improved. John Vandenberg ^(chat) 01:25, 5 June 2008 (UTC)

I had played around with the test wiki a bit and found this new feature to be very complicated. There are three sets of logs to go along with it and none of the actions show up in the page history. Does anyone have firm ideas on how they would like to see it implemented? There are a lot of options, and it will overwrite our current patrolling system. So we should be cautious.--BirgitteSB 01:48, 5 June 2008 (UTC)

We have RC patrol, {{TextQuality}} and {{PageQuality}} which work quite well, so mere quality assessment is not what we need FlaggedRevs for. The new key feature of FlaggedRevs is the ability to hide revisions from anonymous users. A project such as Wikisource probably has much more non-editors ("library users") than editors ("librarians"). We cannot expect most users to know details of the Wikisource QC process, or, for that matter, what a Wiki is. These users deserve to be shown, not the most recent revision, but a library item in acceptable condition, or at least a warning that the condition is not acceptable.

Therefore I'd recommend using FlaggedRevs in a way which complements our established QC process. This means, in particular, that we don't activate all bells and whistles of this extension, with grades from A to F in many categories. Instead I propose the extension to be activated for the main and Page: namespaces only, with stable template transclusion and only two, at most three, grades and no distinction of flag type:

1. checked for vandalism: the revision does not contain obvious vandalism. If this can somehow be linked to the usual RC patrol, we can ditch this grade.
2. acceptable condition: the revision is based on a revision with a quality level of at least 75% and the changes were checked against a reasonably trustworthy source (such as Project Gutenberg, or a recent paperback edition) by a trusted editor.
3. high quality: the revision is based on a revision with a quality level of 100% and the changes have been checked against a scanned original edition available from the Commons (or, in exceptional cases, from Wikisource itself), or, in the case of modern government texts, against the content of an official website, by an admin.

Another advantage of FlaggesRevs would be that we could loosen our protection policy and no longer need to lock texts with 75%+ quality.--GrafZahl (talk) 08:48, 5 June 2008 (UTC)

My concern is that FlaggedRevs can not complement our established QC process, because the extension is incompatible with the existing revision patrolling. Without revision patrolling or some sort of queue of untagged edits, I doubt that we can maintain a high degree of quality new contributions. I care more about expanding our collection, and the quality of our collection, than I care about readers. The readers will come; some will turn into contributors; our credibility will grow. If FlaggedRevs turns off patrolling, it becomes exceedingly difficult to even see new talk comments which need to be answered.

We are in between a small wiki and a large wiki, and revision patrolling is the currently working way to address this.

In short, our RC feed is now too voluminous to be monitored in entirety (without a filter like unpatrolled edits), and we dont yet have enough community involvement to trust people to actively maintain the pages on their watchlist. A year ago, it used to be possible to view all changes by viewing the last 250 recentchanges - in the last 24hrs, Mattwj2002 created more than 250 new pages, and they were all automatically patrolled by JVbot. The day before, the entire 68 pages of The Modern Art of Taming Wild Horses was uploaded by a brand new user Alex brollo.

See our Edits/Day chart from Feb.; I have no doubt that since Feb. the chart is looking even more scary. 2000 revisions occurred in the last 3 days. Solutions which may have seemed fine a year ago may not be suitable now.

In order to stimulate new contributors, we need to carefully inspect their edits and actively help them. We are also going to need to start looking for hoaxes; e.g. these curiosities, which look quite realistic, but dont appear in this edition. Are we going to simply roll them back, and turn away possible new contributors? Investigating these contribs can take days, or weeks, and the unpatrolled queue gives us that capability. Unless FlaggedRevs gives us the same simple "queue" to work on, incoming changes will be left unattended, and new contributors will be turned away because their good edits are kept hidden because they are lost in the history.

I strongly suggest waiting until enWP implements this (their New page patrol is similar to our situation), as they will push to ensure that a workable replacement is bolted onto FlaggedRevs before it is implemented. John Vandenberg ^(chat) 13:14, 5 June 2008 (UTC)

The above comments make it clear that a basic question must be answered before the conversation can continue productively: Is the Flagged Reviews extension incompatible with the existing revision patrolling? What might make it so? No such limitation is mentioned, so what thoughts made people bring up this concern?

Waiting for en.wikipedia might indeed be a good idea to resolve this concern, because if it works for them it might work for us. In general, my suspicion has always been that this feature will reduce a lot of the burden on the people who normally do patrolling. Dovi 13:41, 5 June 2008 (UTC)

I believe people think it is incompatible form the experiences I shared with while testing the test wiki. Also the page reads: $wgUseRCPatrol is enabled with the extension. Flagged revisions are marked as patrolled. Patrolling and autopatrolling of pages in reviewable namespaces is disabled. Other namespaces can be patrolled by Editors. This will mean that the only way to patrol a reviewable revision to to tag it as approved. The advantage is that patrolled edits will always correlate with a reviewed revision. (emphasis mine)--BirgitteSB 16:29, 5 June 2008 (UTC)

It seems my proposal above was premature. First, the general question of practicality raised by John needs to be addressed. I agree that the new extension should not interfere with our established QC process, only improve it. If anything, it may replace parts of our processes with a functionally equivalent feature that is easier to use. With the concerns raised, it's no longer clear to me that this can be achieved with the current version. We would not use the extension for the talk namespaces. Furthermore, it appears the extension has its own autopatrol feature. From Erik's e-Mail: Only changes made by users who are not permitted to patrol changes need to be patrolled. It appears the new extension has the ability to completely subsume RC patrol. But some questions remain: does it work, practically? Does it integrate with the old RC patrol, or do I have to use two different Special pages for reviewable and non-reviewable namespaces? If I want to do RC patrol only, can I tone down the interface to not show me all the fancy higher grading stuff? Can I subsume RC patrol for the whole wiki but activate higher grades only for the wanted namespaces?

BTW, the German page w:de:Hilfe:Gesichtete und geprüfte Versionen does not mention the old RC patrol at all, and on w:de:Hilfe Diskussion:Gesichtete und geprüfte Versionen there is some rather recent lament about lack in user friendliness and some bugs in the extension. So it may really be best to wait a little longer, as Dovi suggests.--GrafZahl (talk) 21:33, 5 June 2008 (UTC)

I agree that it may be worth waiting until en.wikipedia plans what it is going to do about flagged revisions. However I also wonder whether the extension could be implemented at different stages. To begin with we could start by only flagging versions which are currently '100%' or 'validated', and only once this is tested, would we discuss whether it is necessary for the flagging to have more levels and categories. My fear is that if we implement a complex system editors may spend more time flagging revisions, and checking they are up to date, instead of adding new content. Suicidalhamster 22:53, 5 June 2008 (UTC)

This extension is enabled per namespace, so implementing it at all in the main namespace will have an effect on our existing processes for that namespace. I can see this being useful in the "Page" namespace, however that will probably require some adjustment with the Proofread Page extension: i.e. the proofreading flag used in that namespace can probably be replaced with this new technology.

If we wanted to trial the FlaggedRevisions extension in a namespace, in order to get a feel for what it can do, we could pick one or more of the less edited/visited namespaces, such as "Help", "Wikisource", "Author", "Template", and "Index". John Vandenberg ^(chat) 23:47, 5 June 2008 (UTC)

Recent change patrolling is not just about spotting vandalism. It's about noticing a new user who needs welcoming, or spotting innocent typos, or finding a page that needs re-formatting or categories added.--Poetlister 14:29, 6 June 2008 (UTC)

From the very beginning (I think 4th May or so) of the beta running of flagged revisions on the German Wikipedia I took part in it (I asked for the status of so called editor on the first day). I guess it is a good and practicable instrument for wikipedias. But wikisourses have another structure of articles, another structure of users, and another structure of vandalism, therefore another structure of faithing the vandalism. I do not think that this feature could profoundly help here, in contrary, it could make the working here worse. -jkb- (cs.source) 12:48, 8 June 2008 (UTC)

Spurred on by this blog post, I created an account over on en.labs, granted myself "Editor" and "Reviewer" rights, and checked out the Recent Changes.

Surprisingly, I am greeted with the familiar "!" beside many revisions, with "[Mark this page as patrolled]" appearing on a new pages, and "[Mark this change as patrolled]" on a revision.

Also, changes that I made after I granted myself the additional rights are not accompanied with a "!", so it is quite possible that granting those rights is similar to our current system of granting adminship in order to autopatrol the user. Oddly, I am not greeted with an error when I mark my own changes as reviewed (those edits where made before I gave myself additional rights).

So, it seems like my concerns above were based on presumption and the documentation, rather than on reality and personal knowledge. It looks like FlaggedRevs is compatible with our current mechanisms; apologies to all. It would be good if a few others could play around on the test bed to confirm that our current practises will still work with FlaggedRevs. John Vandenberg ^(chat) 01:42, 12 June 2008 (UTC)

Update: The reason why I saw the typical "mark as patrolled" links is because the Template namespace isnt being controlled by the FlaggedRevs extension. However, Recentchanges works in the same manner for all namespaces, irrespective of whether the namespace is using "Revision patrolling" or "Flagged Revisions", so it is still possible to see all recent unpatrolled/unflagged revisions, which is primarily what I was concerned about. John Vandenberg ^(chat) 02:01, 12 June 2008 (UTC)

Global sysyop

There are current plans to implement the Global Sysop feature accross Wikimedia, which would allow a user to have admin powers accross all wikis. The official proposed policy page is at m:Global sysop. There is a poll at m:Metapub#Global sysops (poll) currently underway until June 30 to determine whether the community actually wants this feature. All are welcome to contribute. - Mtmelendez 14:23, 20 June 2008 (UTC)

Hmm, the rejection of this proposal seems very likely at the present time; as we speak, more people are in opposition of this function than are in support. ---- Anonymous Dissident^Talk 15:59, 21 June 2008 (UTC)

It won't affect Wikisource even if passed, as we have several of our own admins.--Poetlister 21:28, 21 June 2008 (UTC)

Other discussions

Composite encyclopedia pages

Eclecticology has created a number of composite encyclopedia articles on Wikisource, which contains copies of articles from various encyclopedias about the subject (such as Kentigern). However, these articles are not part of their main work's namespace and in several cases duplicate existing articles in the proper namespace (such as Munich and 1911 Encyclopædia Britannica/Munich). I originally split the Encyclopedia Britannica articles to their own pages while I was performing bot maintenance, but their contributor reverted.

In my opinion, we should split these composite articles into their respective works, and convert the pages into disambiguation pages. Since these are new to Wikisource, I'm wondering what other contributors think. —{admin} Pathoschild 04:34:29, 20 May 2008 (UTC)

I agree that these need to be split. I have been going through old pages and either splitting them or tagging them with {{split multiple}}. There are some benefits to having the encyclopedic articles considered "works" rather than "subpages" (e.g. if the same article was published in more than one work), however the benefits in having multiple articles on the one page are far outweighed by the limitations in this approach. The most obvious problem is that putting multiple entries in a single page prevents people from only viewing/printing a single entry.

I think I understand what Ec. is trying to achieve, and think there is value in it. It is similar to what has been done with Bible/Jude, where multiple editions of the same Bible verse are displayed together. But, that is an additional layer on top of the Bible translations that are set up to follow the style guidelines. John Vandenberg ^(chat) 07:45, 20 May 2008 (UTC)

I see no reason to stop producing composite encyclopedia pages. Although there has been a clear habit established of treating articles as subpages, there has been no general discussion about why that would be preferable. The force of habit alone does not produce a guideline. Munich did not duplicate 1911 Encyclopædia Britannica/Munich; it was the other way around; the duplicate was more recently extracted from the Munich article.

It is worth considering the use of disambiguation pages, but who is going to do that work? The Catholic Encyclopedia is apparently complete, as is a large part of the first volume of the EB, and we have a number of other encyclopedic works complete as well. Where are the disambiguation pages for these. The argument that users would not be able to see or print a single entry by itself is incredibly weak, but it is the only one that you can come up with. It is much easier for a user to extract the one he wants from many, than to bring together for comparison articles which he may not know to exist.

Also, let's stop putting so much weight on style guidelines. Guidelines are not obligatory. If anyone has a reasonable alternative to the suggested presentation of an article there is no reason for disrespecting those efforts by employing a bully-bot to enforce maniacal standardization. I am willing to respect the work of others, and expect nothing less from those others. Eclecticology 06:40, 21 May 2008 (UTC)

Apart from the specific issue of composite encyclopedia pages which I have no opinion on, I agree with Eclecticology that although some degree of consistency between different works is desirable, there ought to be less insistence on standardization of formatting here at Wikisource compared to the other WikiMedia projects. --❨Ṩtruthious ℬandersnatch❩ 10:10, 21 May 2008 (UTC)

I have no problem with flexibility in the formatting, headers, and the logical structure used to present a work, however Ec. has for six months been adding EB1911 entries in a fashion inconsistent with the main body of our EB1911 work. The time has come to work out what innovations Ec. has come up with, or what problems have been given new light, and re-assert consistency within the EB1911 work.

Also, flexibility in the overall structure of our repository presents a problem. Our "work" definition (i.e. all subpages) is the same used by Wikibooks, and a recently released tool to create a PDF of all subpages should work for Wikisource as well as it does for Wikibooks. Tools depend on standards. Repositories are collections of units that are catalogued. Wikisource currently has no record of how many works are held on Wikisource. John Vandenberg ^(chat) 10:34, 21 May 2008 (UTC)

The previous discussion on this was last edited (I think) December 3, 2007. As far as I can see, this was not archived properly. John Vandenberg ^(chat) 12:56, 21 May 2008 (UTC)

I think it's better to have the text split by work, i.e. on a subpage for each encyclopedia/article. Then, the composite pages could be built up through transclusion, so they could benefit from proofreading on their source pages, etc. -Steve Sanbeg 20:47, 21 May 2008 (UTC)

That would work, but another idea is to have a topic index similar to the indexes in classical encyclopedias. For example:

• Kentigern
  • Collier (1701)
  • Dictionary of National Biography (1885-1900)
  • Encyclopædia Britannica (1911)
  • Catholic Encyclopedia (1913)

—{admin} Pathoschild 02:58:55, 22 May 2008 (UTC)

That was being constructed at Wikisource:Encyclopedia articles, and is a very good idea as the information can then be reused in many ways. For example, if we did decide that we want a disambiguation pages, they could automatically be created by a bot. John Vandenberg ^(chat) 05:40, 22 May 2008 (UTC)

After some thought, I agree with John and Pathos that separate pages is better, for practical reasons. Though I prefer Pathos' suggestion on creating disambig pages for topics, Steve's suggestion of transclusion can be a middle ground we can work with. However, there must be a logical reasoning behind the format we use to present works. Our main objective in this project is to provide sources to the reader, and our format should be user friendly, and direct to the point. So I suggest a having a disambig portion at the beginning of the page (similar to Pathos' example above), because readers may already have an idea on the source they're looking for but haven't found it, followed by the excerpts from major works on the topic (just as Ec has been doing), so that users looking for the information (and not a particular source) may find it easier to research and then cite. - Mtmelendez 15:08, 22 May 2008 (UTC)

I've created an example of my suggestion at User:Mtmelendez/Munich. - Mtmelendez 15:54, 22 May 2008 (UTC)

There are several reasons for the subpage convention we use. A few off the top of my head:

1. The slash-delimited hierarchy is part of the Internet-wide URL standard, so that the convention gives us highly readable, logical, and standard URLs. For example, "en.wikisource.org/wiki/United_States_Code/Title_31/Chapter_69/Section_13" is a much better URL than "en.wikisource.org/wiki/Section_13_%28USC_31/69/13%29".
2. As such, MediaWiki explicitly supports a subpage hierarchy like the one we use. For example, [[../]] on any page links to its parent page, which allows us to use location-independent links. We could rename The Time Machine without needing to make a single link correction to its pages.
3. The resulting URLs are highly stable, since they are based on objective criteria. The page names for The Time Machine will not change in the future based on differing preferences or conditions.
4. The convention makes it easy to find content without searching, so that I know I can find chapter I of The Time Machine at "The Time Machine/Chapter I".
5. The simple standardized hierarchal structure makes Wikisource very easy to contribute to, because it is obvious how pages should be named and organized. (Where should I put Chapter 2 of Work? At "Work/Chapter 2".)
6. The convention makes Wikisource very easy to parse for scripts, which has a whole range of benefits including generating statistics, extracting content automatically, performing services such as printing all pages of the given work, et cetera.
7. The structure makes it very easy to manipulate and duplicate content at will. For example, we could very easily create composite encyclopedia pages that are updated automatically when their source is edited, using simple transclusion.

—{admin} Pathoschild 02:23:03, 23 May 2008 (UTC)

My statement of opinion on this issue became so long and rambling that I decided to just add subheadings and make it a user essay. Maybe the first Wikisource user essay? I notice that it's just lil' ol' me in that category. --❨Ṩtruthious ℬandersnatch❩ 20:04, 22 May 2008 (UTC)

BS, there are two issues here, and your lonely essay appears to have addressed only the one of lesser importance:- the method used to identify a single work. I dont much care how we determine this, except that I have a preference for using "/" because it is' the accepted hierarchical separator in the URL specification, and for the aesthetics benefits of it consequently not being an encoded character (round brackets in URLs do annoy me).

The more important "problem" is pages with text from more than work, which means that the page name is not sufficient to deduce the page contents, and in order to reuse parts of a page, LST is required. This is also another aesthetics problem, because it means that to uniquely identify the text outside of the MediaWiki software, one must then use a URL that links to the section within the page, which makes for longer URLs, and typically results in more encoding in the URL.

John Vandenberg ^(chat) 22:15, 22 May 2008 (UTC)

Okay, just to make sure I'm understanding you here, you're saying that the real problem here is something like, instead of URLs looking like

     http://en.wikisource.org/wiki/1911_Encyclop%C3%A6dia_Britannica/Vol_19_MUN_to_ODDFELLOWS/Munich

they might end up looking like

     http://en.wikisource.org/wiki/Munich#EB1911

I feel as though I must have misinterpreted you but that seems to be a literal example of what you're talking about.

If I am correct that this is what you mean, it does not seem like a substantial problem to me. Not only because the latter URL seems simpler and shorter to me and contains less encoding than the former URL, but because via redirects or even standard transclusion and LST the same content can easily be made to show up under both URLs.

A URL isn't "real", it doesn't have any substantial meaning in and of itself, it's just an address. Just like it says nothing about you as a person whether your postal address includes "street", "road", or "boulevard", a page's URL says nothing about the page - that is the convention of the internet, if anything. If we want certain bits of content to show up when a URL of a certain pattern is typed in to a web browser that's a separate issue from splitting apart the pages that Eclecticology has built.

This isn't something I'm fanatic about and it doesn't seem to affect any of the work I'm doing at the moment. If you guys honestly feel that what I'm saying does not address the real issues here, go ahead with how you were handling this. But I'm saying that the reasoning as presented for taking apart Eclecticology's articles does not seem sound to me and for the record's sake I hope I have thoroughly explained why I do not concur with it. --❨Ṩtruthious ℬandersnatch❩ 15:05, 23 May 2008 (UTC)

Look... to the degree that I have interacted with him so far, it appears to me that Eclecticology is an abrasive jerk. So I can sympathize with any desire to put him in his place. But it also appears that he is a substantial contributor to this project and reasonable requests from him should be accorded flexibility and deference.

He isn't being disruptive here. The worst that's happening is that he's insisting that the composite encyclopedia articles should be the primary presentation of some works when perhaps they ought to only be secondary presentation. As far as I can see it's no sweat to break out the LST and transclude his stuff into pages under contiguous-work-hierarchy-consistent URLs. Yes, this may bear a future cost in increased sophistication on the tech side of things but that's just not the trump card it's being played as and many of the accompanying counterarguments are looking extremely contrived. --❨Ṩtruthious ℬandersnatch❩ 15:47, 23 May 2008 (UTC)

The main argument in favour of his method seems to be that active contributors should do whatever they want to regardless of how inconvenient it is to everyone else. However, it would be better (for all the reasons listed above) to place these works in their standard locations and formatting, and transclude their content onto his composite encyclopedia pages. This is very easy to do, simply a matter of adding "{{Encyclopædia Britannica/Munich}}" to the composite page. It is far messier and more fragile to perform the reverse, extracting the content from a merged page to its standard location.

This is not simply a question of whether Eclecticology should do this, but whether it's fine for any editor to add works with any naming convention and format they wish, duplicate existing entries, and generally make a mess. —{admin} Pathoschild 23:31:48, 23 May 2008 (UTC)

Calling what he's doing "generally making a mess" is purely pejorative and no better an argument than anything else that has been put forward. You're saying that even though both instances of composite pages would look exactly the same to a web visitor, and the only real difference is the location of transclusion code, your version of how to arrange things is virtuous and orderly, whereas his version is a "mess"? Surely you must realize how emptily polemical that sounds. If a work's conventionally-named page can be made up entirely of transclusion code pointing to Page: namespace transcription project elements, the way works originating in a .djvu do, without being a "general mess" it can just as easily, and in an equally-orderly fashion, be made up of transclusion links pointing to identically-named composite encyclopedia articles.

This is splitting hairs. It's not fragile in either case because there's an explicit, parseable transclusion link that a bot could handle in exactly the same way MediaWiki handles it. If you're writing code that makes guesses based upon a supposed URL-expressed structure of the site, rather than parsing the transclusion code to find out where content is coming from, that is what's fragile, not any particular arrangement of pages.

If you're concerned that having composite encyclopedia pages will result in the entire superstructure of Wikisource dissolving into chaos, how about we discuss some way to encode within policy what Eclecticology is doing? By my reading he wasn't exactly reckless about it, he very gradually, openly, and with respect to the work of others has been doing this in an orderly fashion. Like I said, I don't have any problem with placing restrictions on people, it's just that the presented reasoning behind trying to restrict Eclecticology in this case is not sound. --❨Ṩtruthious ℬandersnatch❩ 16:58, 25 May 2008 (UTC)

I think there are two conflicts here, philosophical and technical.

• Philosophical: I first joined Wikisource when it was in a state of acute disorganization, with no template or style standardization whatsoever, and missing most of its structure. It took several years of painstaking effort on the part of the entire community to unify everything into a single standard, unify the templates, and build the structure. With that Sisyphean task finally accomplished after several years, I'm very reticent about turning around and encouraging destandardization.

  In contrast, Eclecticology left the project long before this massive standardization and structure-development effort. He only rejoined recently, to find the project completely different and standardized. He holds standardization in high contempt (for example, he mentions "employing a bully-bot to enforce maniacal standardization"). He's clearly very open to destandardization, and the composite encyclopedia pages are not the only place he has gone against established standards or practice.

• Technical: The technical details are less subjective. The composite pages, if we choose to keep them, can be generated in two ways:

Transcluding standard to composite pages
text	→	{{Encyclopædia Britannica/Munich}}
Transcluding composite to standard pages
<section begin="EB1911" />text<section end="EB1911" />	→	{{#lst:Munich|EB1911}}

As shown above, it's much simpler, less fragile, and less confusing to transclude from standard to composite pages. Consistency is also important for the 1911 Encyclopædia Britannica, which already has thousands of standard pages added; it's a bad idea to transition from one format to another in the same work, and also a bad idea to have duplicate entries. There is also the significant possibility that an editor not familiar with the labeled section transclusion extension will be confused by the syntax, unable to deduce where the content is coming from (There's no "#lst:Munich" page!), mistakenly break the transclusion by changing to a better section name, et cetera.

However, Eclecticology has already shown some interest in using standard disambiguation or index pages instead. There is already a start at Wikisource:Encyclopedia articles. That might be a compromise worth exploring. —{admin} Pathoschild 20:11:29, 26 May 2008 (UTC)

I certainly agree that pursuing a compromise would be a worthwhile effort. And I would encourage Eclecticology to cooperate in efforts towards standardization for standardization's sake - I just don't think, in this particular instance, he's doing anything that indicates he needs to be compelled to cooperate.

The possibility that inexperienced editors might screw things up in the future is a good reason to provide better instructions or a better user interface, or even to have a bot go and place inline comments next to instances of LST usage to explain how to work with them. But it's not a sound justification for taking any deliberately overriding actions against Eclecticology's work here.

The earlier part of this discussion proposed splitting up the composite articles, presumably via Pathoschild's approach of deleting portions of them and creating new articles from the deleted text. But I hope we can all agree that no reasonable cause for doing that has been articulated. Eclecticology, would you have any objections if your composite articles were left so as to maintain the current appearance to web visitors, but the underlying code was shifted around so that everything is being transcluded from articles under orthodox URLs, as in Pathoschild's "standard to composite" example immediately above? --❨Ṩtruthious ℬandersnatch❩ 18:40, 27 May 2008 (UTC)

Thank you for the support from an unexpected source. This is a possible solution, and it would not be appropriate to dismiss it out of hand. Before giving a positive response, however, I would like to see how such a page would look, and how it would be edited to add articles from other sources. Eclecticology 09:45, 30 May 2008 (UTC)

As a demonstration I have converted the EB1911 portion of Münsterberg to be transcluded.

The procedure would be:

1.

You (or whoever's rigging up the transclusion) would create the individual encyclopedia articles under the orthodox URLs just as you would normally, with the exception that you must explicitly set the section=... parameter of the header to the name of the article.

2.

Then in the composite article each individual article can be brought in with code like {{:1911 Encyclopædia Britannica/Münsterberg}}, which as you can see causes the individual article to behave as a template.

--❨Ṩtruthious ℬandersnatch❩ 00:54, 31 May 2008 (UTC)

I read Struthious's essay with interest, and, unlike him, John and Pathoschild I do not come from a software engineering background. This makes me resistant to technical solutions which often strike me as opaque and unduly restrictive. I very strongly adhere to the wiki principle that editing should be easy for everyone. Those editors include many whose understanding relates more to content than format, and a lot of excessive markup will deter many from editing.

In his examples in support of a subpage structure Pathoschild uses two examples, The Time Machine and The United States Code. These lend themselves well to that kind of structure. A novel involves a series of chapters that normally require reading in a certain orderly fashion. A legal code depends on a predetermined structure for its coherence. But the validity of the structure should not depend on the simple easy examples. Eclecticology 09:45, 30 May 2008 (UTC)

I especially appreciate Struthious's recognition of my point about multiple hierarchies. This is key to making knowledge useful. In encyclopedic works we need an easy way to link between the same subject in different encyclopedias. The Britannica is well known, but its various editions had numerous general contemporaries, and there are even more specialized works such as gazeteers and bigraphical that may give different or greater details within their terms of reference. While some of us in our younger days may have wanted to read our encyclopedias from cover to cover, I think that we would most commonly seek information on a specific subject. Our recommended way of organizing this kind of material does not lend itself easily to that kind of search.

The problems of hierarchization are not limited to encyclopedic articles. Novels, like The Time Machine are easily put into hierarchical schemes. Even when they were originally serialized in a magazine, as was frequently the case in the 19th century, they retained an orderly thread between the episodes. On the other hand short stories and poems remain independent works, and their subsequent inclusion in collections does not otherwise create a link between successive works in a collection. It often signals little more than the convenience of the accumulating editor. We have already had an inconclusive discussion about obituaries; a similar problem exists with reviews. Speaking only of works where I have hard-copy material, I shudder to think of how we could best handle Johnson's dictionary or Poole's Index to Periodical Literature. We really need to avoid being short-sighted.

The page Wikisource:Encyclopedia articles is a fine idea, except for the fact that nobody is using it. An important proof of any scheme is whether anyone uses it. We have no end to idiosyncratic ideas that are never applied except by the person who invented the particular scheme, and even then that individual often quietly stops using it when he tires of it. Idiosyncratic schemes are still important because they can become the nucleus of future innovation. The history of invention is full of ideas that never got anywhere; one only needs to look at the records of the patent office to find them. Premature standardization that tries to catch everything only makes innovation more difficult. If an idea bears the hallmarks of abandonment, that is perhaps the time to reabsorb the material into a more standard presentation. Workable, but different, ideas that do not conform to guidelines should never be discouraged. Perhaps the guidelines should be easily amended to show the alternative approaches that are still in play. Let's remember that guidelines are not policies to be enforced.

Our attempts to categorize articles has been non-existent. Understandably, this is difficult work. It requires some analysis of the article's content. We have a scheme to categorize authors, but that frankly does not give a lot of benefit. Most authors have Wikipedia articles, and the few who are missing theres probably should have. That project is in a much better position to categorize authors. For Wikisource the important content is in what those authors wrote, not the authors themselves.

As I pointed out in #Encyclopedic articles below, I did uncover a number of issues around the naming of articles, notably what to do when the same article is named differently in two different publications. Perhaps another kind of standardization to consider is standard titles that would transcend the variations that can and do occur. Thus the article title would show the standard title, but the actual article title would appear in the header.

In summary, I very strongly believe that the software should serve the content, not the other way around. Also, many attempts to simplify templates really end up being simplistic. Eclecticology 18:31, 30 May 2008 (UTC)

:P--T. Mazzei 04:17, 23 May 2008 (UTC)

Done. —{admin} Pathoschild 06:20:55, 23 May 2008 (UTC)

In part. Eclecticology 18:31, 30 May 2008 (UTC)

While I consider the effort by Struthious to find a solution a step in the right direction, I am somewhat disappointed by the lack of response to my comments. We need to be able to deal with these kinds of issues with considerably more foresight than has been the case, without trying to shoehorn everything into a one size fits all model of templates. Eclecticology 06:29, 18 June 2008 (UTC)

Questions

Sufism/Nasrudin

User:Mike.lifeguard recently moved content here from Wikibooks that some thought would better belong on this project. The material is up for deletion, but consensus leans towards keep. It might, however, be deleted if Wikisource is a better place for it.

These are folk-tales that have been published, but I don't believe we have any sources as to which publication (if any) these are from. I've consulted Wikisource:What Wikisource includes and Wikisource:What is Wikisource? but I'm still not sure about whether this is appropriate for Wikisource. --Swift 02:26, 11 June 2008 (UTC)

I cannot find any source suggesting this was published in a "real book", which we require, and just a few blogs that have copied it. If it was written in the last decade by online authors - then it does not fit Wikisource criteria I'm afraid. Sherurcij ^{Collaboration of the Week: Author:John McCain and Author:Barack Obama} 02:43, 11 June 2008 (UTC)

A quick internet search reveals among others the Wikipedia article w:Nasreddin which lists five collections of these books. These are folk stories, not written by any one person, but have lived through retelling. There are no sources as to where the contributors got their versions nor to what extent they added or molded the story.

I guess the question is whether the requirement is that the text as it appears on Wikisource must have been published somewhere, or just the storyline. In the case of folk-tales there is, by definition, no identifiable source. --Swift 13:50, 11 June 2008 (UTC)

Folk-tales do have an identifiable sources even if they do not have an indentifiable single point of origin. Also translators of such tales have their own copyright on the translation. At least we need an idenfiable source for the English version with an acceptable copyright status to keep these here.--BirgitteSB 16:11, 11 June 2008 (UTC)

OK, I think that settles it. I haven't seen any identifiable source for the text so I fear you may have to delete Sufism/Nasrudin. Thank you very much for your help. --Swift 15:40, 12 June 2008 (UTC)

Available classes for HTML elements

Where can I find out about available classes for HTML elements, such as

• lefttext
• prose
• editsection

and the like? Is there a guideline? I had a look at various help pages here, without success. --Dan Polansky 13:54, 14 June 2008 (UTC)

Duplicate, I think

I think we have two versions of the code of Hammurabi (The Code of Hammurabi and Codex Hammurabi). I'm really not sure what to do, which is why I'm posting it here. Psychless 22:21, 14 June 2008 (UTC)

Thanks, I redirected both to their respective translators names - and made The Code of Hammurabi a choice between the two - similar to what we do with things like Genesis or The Confessions of Saint Augustine Sherurcij ^{Collaboration of the Week: Author:John McCain and Author:Barack Obama} 23:00, 14 June 2008 (UTC)

Thanks. I fixed the link at Author:Hammurabi. Psychless 02:02, 15 June 2008 (UTC)

DNB00 project?

Hi, all. Is there a project to add the 1900 edition of the Dictionary of National Biography to Wikisource? Isee several articles that appear to belong to the "project," but I cannot find the project itself. See This search. The pages sort of follow the format of the 1911EB project. -Arch dude 15:05, 15 June 2008 (UTC)

It is currently a project with only one person, Eclecticology, who would love to have some assistance. I have created Dictionary of National Biography, 1885-1900 so that it has a talk page where discussion can commence. John Vandenberg ^(chat) 05:36, 16 June 2008 (UTC)

Thanks. Eclecticology 09:09, 17 June 2008 (UTC)

OK, I created a "stub" project by slavishly copying and crudely adapting the EB1911 project. I probably made a bunch of mistakes, so please be bold and fix them as you see fit. -Arch dude 02:58, 17 June 2008 (UTC)

I'll see you there. Eclecticology 09:09, 17 June 2008 (UTC)

New South African Constitution

I'm really unsure as to where to ask this. I've searched Wikisource extensively, including the forum, but can't see that this has been discussed before.

I'm wondering if the Constitution of South Africa (1996) can be uploaded here. The text is available from a government website and the Constitutional Court, amongst other pages. However, both these sites disallow reproduction for commercial use, which I presume prevents reproduction on Wikisource. Warrickball 18:25, 16 June 2008 (UTC)

On the other hand the US does not recognize attempts to copyright laws (Although Oregon seems to have forgotten this recently). {{PD-EdictGov}}--BirgitteSB 02:44, 17 June 2008 (UTC)

You may want to see w:Official_text_copyright#South_Africa. I agree BirgitteSB's opinion while we already have some non-USA governmental edicts after extensive arguments. However, disallowing copyright in the USA does not necessarily affect the copyright in the source country, so please be aware.--Jusjih 02:59, 17 June 2008 (UTC)

Yes if you are not in the US yourself. Be aware of local laws and any possible personal liability.--BirgitteSB 03:05, 17 June 2008 (UTC)

An interesting question I'm sure hasn't yet come before the courts, on Wikisponsibility...I can add the document, even though he can't - but can he then edit/improve the text, or would clicking "Save" be considered publishing a (revised) copy of the entirety? Sherurcij ^{Collaboration of the Week: Author:John McCain and Author:Barack Obama} 03:24, 17 June 2008 (UTC)

As always, this would depend on the jurisdiction of the copyright holder and of the contributor, and I would not like to see Wikisource develop policies or guidelines that prevent contributors from making certain edits. At most we should have a project page which explains that contributors may be at risk if they edit certain pages, but that risk needs to be assessed and managed by the contributor. For example, I have uploaded and edited many {{PD-1923}} texts which are probably not PD in Australia, simply because I don't believe that the edits put me at significant risk.

What should be pointed out is that the nationality of any editor is not guaranteed to remain private, so it is the contributors responsibility to keep their identity and nationality well hidden if they are doing something that puts them at unacceptable risk. John Vandenberg ^(chat) 04:06, 17 June 2008 (UTC)

Thanks for that link, Jusjih. It appears that there is no copyright on "official texts of a legislative, administrative or legal nature". I presume the Constitution is at least one of these, so I'll go ahead and upload sometime. Thanks for the speedy and helpful responses! Warrickball 10:10, 17 June 2008 (UTC)

20 MB upload limit on the Commons

There is currently a 20 MB upload limit on the Commons. This causes problems for us to upload large djvu files. There is currently bug on this issue. This bug can be found here. Please share your opinion here and on the bug. I think it is important that we don't have to split our books into smaller files. --Mattwj2002 23:13, 18 June 2008 (UTC)

I'm pretty sure that spamming the bug report with "I agree; please fix it" will only piss off the developers. The correct process is to discuss it on wiki, then, once consensus has been achieved, submit a bug report/update linking to the discussion. Hesperian 23:29, 18 June 2008 (UTC)

Yes, agree with Hesperian that we should discuss here first... and so to start things off I agree that the limit should be raised somewhat. giggy _(:O) 01:58, 19 June 2008 (UTC)

So do I :) - --Zephyrus 21:22, 19 June 2008 (UTC)

-> Hesperian: Well, there is already a bug report: no need to submit another one. But you can vote for the bug. Yann 21:35, 19 June 2008 (UTC)

KAL 007

Not sure of the right place to put this; please move it if it's not right. Could an admin please transwiki w:KAL_007:_Inside_the_Cockpit and let me know when it's done so it can be deleted, or alternatively let me know if it's not suitable for transwiki to Wikisource. Best way to get hold of me is here. Thanks. Neil 13:10, 19 June 2008 (UTC)

It has already been deemed unsuitable[1]; see w:Talk:Korean Air Lines Flight 007#related wikisource text and w:Talk:Korean_Air_Lines_Flight_007#moved_to_Wikisource. John Vandenberg ^(chat) 21:25, 19 June 2008 (UTC)

Global deleted image review

FYI, a discussion about allowing commons sysops the right to view deleted images on any wikimedia project is happening on Meta here: m:Metapub#Global_deleted_image_review. giggy _(:O) 07:04, 23 June 2008 (UTC)

Links to Wikisource in other languages

Hi, if we have the same work in more than one translation, should all the translations link to other language wikisources? The Confessions of Saint Augustine is currently a disambiguation page with links to two different translations - The Confessions of Saint Augustine (Outler) and The Confessions of Saint Augustine (Pilkington). The disambiguation page does not link to the original, and neither does the Pilkington translation, but the Outler translation does. What's the proper way of doing it? Thanks. Stratford490 12:25, 23 June 2008 (UTC)

You're correct that all three translations should point to all other languages (If for example, there were two German translations, we would link them to the German disambig page). I've done it for Confessions just now. Thanks! Sherurcij ^{Collaboration of the Week: Author:John McCain and Author:Barack Obama} 12:36, 23 June 2008 (UTC)

Thanks, and sorry for giving the wrong link. As I see you discovered, I meant this one. My Latin isn't actually that bad. I was looking at both pages, and just copied and pasted the wrong URL! Stratford490 13:23, 23 June 2008 (UTC)

Yeah, I just copy/pasted your link and then got up to make dinner. Standing at the stove it suddenly hit me "Wait a second...that was City of God..." and I returned to the computer to fix my error ;) Sherurcij ^{Collaboration of the Week: Author:John McCain and Author:Barack Obama} 20:15, 23 June 2008 (UTC)

Obviously we both know enough Latin to be able to work out which is which if we have "Confessiones" and "De civitate Dei" and know that one is "Confessions" and the other is "City of God". I'm not at all sure that I'd have been able for it if it had been Chinese! Thanks for your help. Stratford490 20:38, 23 June 2008 (UTC)

Integrating search engines such as Google

What about integrating search engines such as Google into the search function of Wikisource, modeled on Wikipedia[2] and Wiktionary[3]? See the linked pages and the combo box defaulting to the option "MediaWiki Search" --Dan Polansky 13:01, 23 June 2008 (UTC)

I think it is a good idea. Do you know how it was done there?--BirgitteSB 11:12, 25 June 2008 (UTC)

It goes in MediaWiki:Common.js. Look at wikt:MediaWiki:Common.js, and a Ctrl+F for "Change Special:Search to use a drop-down menu" - copying everything from there to the bottom into our Common.js should do it. giggy _(:O) 11:50, 25 June 2008 (UTC)

As our version of the page is far from blank, I don't feel competent to make any changes there.--BirgitteSB 14:38, 28 June 2008 (UTC)

Printable Version

From the WP Signpost:

A new system message MediaWiki:Print.css allows administrators to edit the print stylesheets (used for the 'printable version', and automatically when printing in several browsers). (r36308, bug 2889)

I think this might be a feature we would want to make use of, but I am not sure exactly how much can be done with the "print stylesheets". Does anyone more technical have ideas?--BirgitteSB 11:12, 25 June 2008 (UTC)

This is the page it affects (taking the example for the Scriptorium). I suppose any sort of CSS can be added to pretty it up, though I don't know how many people actually use printable versions (it could be a good number on this project, come to think of it) and what needs prettying. giggy _(:O) 11:44, 25 June 2008 (UTC)

Help

Coming of Age in Samoa, a classic 1928 text by w:Margaret Mead. Stanford copyright database appears to list renewals on two prefaces and a foreward, but not on the text itself. So I'm thinking it can be added, yes? Sherurcij ^{Collaboration of the Week: Author:John McCain and Author:Barack Obama} 21:25, 25 June 2008 (UTC)

No. For R155003 the "long record" shows this as Mead's own renewal for her text within the appropriate time frame. She would not have had the right to renew the Boas foreword, so at least the foreword has likely fallen into the public domain. Eclecticology 19:42, 26 June 2008 (UTC)

Constitution of the Principality of Sealand

An IP has raised the issue of the above noted text being a fake, or otherwise outdated. See Talk:Constitution of the Principality of Sealand. What are people's thoughts on this? —giggy 15:21, 27 June 2008 (UTC)

No source has been given for the document that we have. Where was it published? While legitimate constitutional documents may be treated as not copyrightable, in the absence of specific evidence that the author has put the document into the public domain, the text could be a copyright violation. If there is a dispute about which text is the valid one, the competing versions should receive equal treatment. Eclecticology 19:04, 27 June 2008 (UTC)

Do ut des: cross validation

I finished a couple of horsemanship-related books, but I can't to the final step by myself... t.i. validation of Index pages. While waiting for some volunteer, I suddendly realized that anyone, working with Index versions, has exactly the same problem.... to find some other interested volunteer for such a boring work. And I realized that I can't hope to find such a volunteer easily.... if I am not ready and happy to validate Index works of some other user!

So, can we organize some special page to list do ut des volunteers giving mutual help and assistance - t.i. promising to validate a work of a user, and getting the promise that the same user will validate their work?

Here my works in need of validation:

Index:The Modern Art of Taming Wild Horses.djvu 1858, a brief book bu J. S. Ramey, mostly interesting for any rider and particularly to those interestin about "horse whispering";

Index:Equitation.djvu, 1922, a comprehensive book by H. L. de Bussigny, a French high level rider previouslòy student under Comte D'Aure and deeply bond with Baucher and Fillis methods.

I'd like to validate any work dealing with nature or science in general. --Alex brollo 10:21, 29 June 2008 (UTC)

Meanwhile, you might try User:Quadell/Proofing swap meet. Hesperian 11:37, 29 June 2008 (UTC)

Ok. I simply "ho scoperto l'acqua calda" (I discovered the warm water, as we say in Italy). Thanks for the link. ;-) --Alex brollo 13:19, 29 June 2008 (UTC)

Sahih Bukhari

I came across Sahih Bukhari, and I noticed it needs serious work. Is anyone comfortable working with texts about Islam? The pages probably need to be moved to Book X subpages and formatted, and some edition/translator info needs to be found. Most of the work on the text was done by IPs. Psychless 16:28, 30 June 2008 (UTC)