Recap: A while ago we discussed date conditional switching templates: http://lists.wikimedia.org/pipermail/wikien-l/2009-May/100714.html . The problem to be corrected was the use of future tense language which then becomes outdated and thus notably incorrect. This also has a greater effect of casual correction patterns which essentially annotate the error rather than fixing it. For example:
""Apple's iTunes store *will start* to sell DRM-free 256 kbit/s (up from 128 kbit/s) AAC encoded music from EMI for a premium price (this has since reverted to the standard price).""
A proper correction would have simply changed "will start [to sell]" to "began [to sell]" and that would be that. Time and tenses require a little bit of thinking however, and an editor made a parenthetical comment (edit note, annote) in place of a considered switch of tense. Forgivable but incorrect. If the {{dateswitch}} template idea was fully implemented and used, anyone writing future events could simply write {{dateswitch|will start|began|ON DATE}} and the switch would happen on the date.
The idea had some support, but people had some issues with dateswitch templates that would produce the wrong output because of some later change in the input. I guess that this might be more rare than common. The above example is notable however of where they miss the point. I note that we now have a category for some tags which relate to time, but I don't know about some of them: http://en.wikipedia.org/wiki/Category:Temporal_templates These appear to be largely template messages, and if we are to employ actual computational power in helping deal with outdating, would it make sense to make a distinction between temporal messages and temporal (functional) tags?
-SC
On Tue, May 18, 2010 at 10:24 PM, stevertigo stvrtg@gmail.com wrote:
Recap: A while ago we discussed date conditional switching templates: http://lists.wikimedia.org/pipermail/wikien-l/2009-May/100714.html . The problem to be corrected was the use of future tense language which then becomes outdated and thus notably incorrect. This also has a greater effect of casual correction patterns which essentially annotate the error rather than fixing it. For example:
""Apple's iTunes store *will start* to sell DRM-free 256 kbit/s (up from 128 kbit/s) AAC encoded music from EMI for a premium price (this has since reverted to the standard price).""
A proper correction would have simply changed "will start [to sell]" to "began [to sell]" and that would be that. Time and tenses require a little bit of thinking however, and an editor made a parenthetical comment (edit note, annote) in place of a considered switch of tense. Forgivable but incorrect. If the {{dateswitch}} template idea was fully implemented and used, anyone writing future events could simply write {{dateswitch|will start|began|ON DATE}} and the switch would happen on the date.
The idea had some support, but people had some issues with dateswitch templates that would produce the wrong output because of some later change in the input. I guess that this might be more rare than common. The above example is notable however of where they miss the point. I note that we now have a category for some tags which relate to time, but I don't know about some of them: http://en.wikipedia.org/wiki/Category:Temporal_templates These appear to be largely template messages, and if we are to employ actual computational power in helping deal with outdating, would it make sense to make a distinction between temporal messages and temporal (functional) tags?
-SC
Could a temporal template in one way or another subst: itself once the expiriation date passes? Now that would be awesome.
Martijn
A self-updating template seems very unlikely to this non-programmer, but a bot could certainly be created to handle these tasks. I wonder if that isn't already done on, say, {{as of}} for example?
On Tue, May 18, 2010 at 5:12 PM, Martijn Hoekstra <martijnhoekstra@gmail.com
wrote:
On Tue, May 18, 2010 at 10:24 PM, stevertigo stvrtg@gmail.com wrote:
Recap: A while ago we discussed date conditional switching templates: http://lists.wikimedia.org/pipermail/wikien-l/2009-May/100714.html . The problem to be corrected was the use of future tense language which then becomes outdated and thus notably incorrect. This also has a greater effect of casual correction patterns which essentially annotate the error rather than fixing it. For example:
""Apple's iTunes store *will start* to sell DRM-free 256 kbit/s (up from 128 kbit/s) AAC encoded music from EMI for a premium price (this has since reverted to the standard price).""
A proper correction would have simply changed "will start [to sell]" to "began [to sell]" and that would be that. Time and tenses require a little bit of thinking however, and an editor made a parenthetical comment (edit note, annote) in place of a considered switch of tense. Forgivable but incorrect. If the {{dateswitch}} template idea was fully implemented and used, anyone writing future events could simply write {{dateswitch|will start|began|ON DATE}} and the switch would happen on the date.
The idea had some support, but people had some issues with dateswitch templates that would produce the wrong output because of some later change in the input. I guess that this might be more rare than common. The above example is notable however of where they miss the point. I note that we now have a category for some tags which relate to time, but I don't know about some of them: http://en.wikipedia.org/wiki/Category:Temporal_templates These appear to be largely template messages, and if we are to employ actual computational power in helping deal with outdating, would it make sense to make a distinction between temporal messages and temporal (functional) tags?
-SC
Could a temporal template in one way or another subst: itself once the expiriation date passes? Now that would be awesome.
Martijn
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: https://lists.wikimedia.org/mailman/listinfo/wikien-l
Martijn Hoekstra <martijnhoekstra@gmail.com
Could a temporal template in one way or another subst: itself once the expiriation date passes? Now that would be awesome.
William Beutler williambeutler@gmail.com wrote:
A self-updating template seems very unlikely to this non-programmer, but a bot could certainly be created to handle these tasks. I wonder if that isn't already done on, say, {{as of}} for example?
IANAP so take with a grain of salt, but this is my sense of things. There are bots that work with templates, categories, etc.. AIUI, for this kind of template to work requires a switching function that would be based on the conditional {{if}} templates. One of the complaints about the template system is that things can only be done through these boolean templates, which locks the programmer into certain ways of doing things which themselves are not objectively functional, for example searching the database. Combining template functions with bot processes is something people have done, but AFAIK noone has worked on developing this combination as an actual computing paradigm. During debates about implenting a scripting language, people seemed hung up on the surface-level issues with template scripting.
-SC
On 18 May 2010 21:24, stevertigo stvrtg@gmail.com wrote:
Recap: A while ago we discussed date conditional switching templates: http://lists.wikimedia.org/pipermail/wikien-l/2009-May/100714.html . The problem to be corrected was the use of future tense language which then becomes outdated and thus notably incorrect. This also has a greater effect of casual correction patterns which essentially annotate the error rather than fixing it. For example:
""Apple's iTunes store *will start* to sell DRM-free 256 kbit/s (up from 128 kbit/s) AAC encoded music from EMI for a premium price (this has since reverted to the standard price).""
A proper correction would have simply changed "will start [to sell]" to "began [to sell]" and that would be that. Time and tenses require a little bit of thinking however, and an editor made a parenthetical comment (edit note, annote) in place of a considered switch of tense. Forgivable but incorrect. If the {{dateswitch}} template idea was fully implemented and used, anyone writing future events could simply write {{dateswitch|will start|began|ON DATE}} and the switch would happen on the date.
[[WP:V]] says no. The use of the wrong tense has the additional benefit in that it instantly indicates that there is something that needs updating here.
geni geniice@gmail.com wrote:
[[WP:V]] says no. The use of the wrong tense has the additional benefit in that it instantly indicates that there is something that needs updating here.
You assume that such errors are good, just because they are obvious. While I respect your uberterseness, you miss the point that these are errors which are not always obvious and therefore go undetected. You also miss the point that "date" here refers not just to tenses, but to any context where language refers to future events. That such events often do not come is one objection, but it does not to my mind rise to become an issue of V.
So what in V would prohibit date-based tense or context switches?
-SC
On 19 May 2010 00:04, stevertigo stvrtg@gmail.com wrote:
geni geniice@gmail.com wrote:
[[WP:V]] says no. The use of the wrong tense has the additional benefit in that it instantly indicates that there is something that needs updating here.
You assume that such errors are good, just because they are obvious. While I respect your uberterseness, you miss the point that these are errors which are not always obvious and therefore go undetected. You also miss the point that "date" here refers not just to tenses, but to any context where language refers to future events. That such events often do not come is one objection, but it does not to my mind rise to become an issue of V.
So what in V would prohibit date-based tense or context switches?
-SC
Because at best you have a source that something is expected to happen at a future date. you do not have a source saying it actually has happened and thus a tense change leads to misrepresenting the source.
geni geniice@gmail.com wrote:
Because at best you have a source that something is expected to happen at a future date. you do not have a source saying it actually has happened and thus a tense change leads to misrepresenting the source.
Well put. So your concern is that outdated phrases and claims be properly corroborated with their respective outdated sources? Would you suggest that an automagic solution for both problems is impossible?
-SC
On Wed, May 19, 2010 at 1:58 AM, stevertigo stvrtg@gmail.com wrote:
geni geniice@gmail.com wrote:
Because at best you have a source that something is expected to happen at a future date. you do not have a source saying it actually has happened and thus a tense change leads to misrepresenting the source.
Well put. So your concern is that outdated phrases and claims be properly corroborated with their respective outdated sources? Would you suggest that an automagic solution for both problems is impossible?
The usual solution is to use templates to flag up outdated material for editors to review and check things did happen, and to issue the needed corrections and updates.
Carcharoth
Carcharoth carcharothwp@googlemail.com wrote:
The usual solution is to use templates to flag up outdated material for editors to review and check things did happen, and to issue the needed corrections and updates.
Exactly. Its quaint. Tagging specific forecasts as temporal takes what is a lazy eventualistic, let-someone-else-do-it process and signals that specific outdated language (and its outdated reference) need updating. A "switchbot" could then perhaps message the editor who wrote the section/tag. Context-based signalling.
-SC
On Wed, May 19, 2010 at 2:39 AM, stevertigo stvrtg@gmail.com wrote:
Carcharoth carcharothwp@googlemail.com wrote:
The usual solution is to use templates to flag up outdated material for editors to review and check things did happen, and to issue the needed corrections and updates.
Exactly. Its quaint. Tagging specific forecasts as temporal takes what is a lazy eventualistic, let-someone-else-do-it process and signals that specific outdated language (and its outdated reference) need updating. A "switchbot" could then perhaps message the editor who wrote the section/tag. Context-based signalling.
The "as of" system (when I last looked) did put articles in a category automatically. Your idea of signalling the original editor is an interesting one, but editors come and go, and the whole wiki-model relies on people updating the work of others, so we need to make sure that process always works.
Carcharoth
On Tue, May 18, 2010 at 6:39 PM, stevertigo stvrtg@gmail.com wrote:
Carcharoth carcharothwp@googlemail.com wrote:
The usual solution is to use templates to flag up outdated material for editors to review and check things did happen, and to issue the needed corrections and updates.
Exactly. Its quaint. Tagging specific forecasts as temporal takes what is a lazy eventualistic, let-someone-else-do-it process and signals that specific outdated language (and its outdated reference) need updating.
Wikis are all about eventualistic processes. It's kind of the point.
That said, automatic notification of people who care that a statement has become stale is useful. Perhaps this could set a special flag in watchlists?
I'm against the idea of automatically updating content for the reason that it's assuming the prediction in the article about a future event is true and actually happened as predicted.
-Matt
Matthew Brown morven@gmail.com wrote:
Wikis are all about eventualistic processes. It's kind of the point. That said, automatic notification of people who care that a statement has become stale is useful. Perhaps this could set a special flag in watchlists? I'm against the idea of automatically updating content for the reason that it's assuming the prediction in the article about a future event is true and actually happened as predicted.
That's an interesting compromise.
-SC
On Wed, May 19, 2010 at 4:09 AM, stevertigo stvrtg@gmail.com wrote:
Matthew Brown morven@gmail.com wrote:
Wikis are all about eventualistic processes. It's kind of the point. That said, automatic notification of people who care that a statement has become stale is useful. Perhaps this could set a special flag in watchlists? I'm against the idea of automatically updating content for the reason that it's assuming the prediction in the article about a future event is true and actually happened as predicted.
That's an interesting compromise.
It's actually a refutation of your position (as were all the other posts pointing out that automatic updating is a bad idea), but if you want to call it a compromise, fair enough. :-)
Carcharoth
Just remembered that we do have time-dependent updating in the "age" bits of infoboxes. People's ages increase by one each year and they are assumed to be still alive until such time as their death is reported or they look a bit too old. But that is the only case I am aware of where content is automatically updated over time. But there may be others, of course.
Carcharoth
Carcharoth carcharothwp@googlemail.com wrote:
Just remembered that we do have time-dependent updating in the "age" bits of infoboxes. People's ages increase by one each year and they are assumed to be still alive until such time as their death is reported or they look a bit too old. But that is the only case I am aware of where content is automatically updated over time. But there may be others, of course.
So where its a simple matter of math, there's no verifiability issue. There is a general substantive argument against automated content creation, and most objections sit in that context. Age calculation could be argued is an example which violates the principle. Its interesting: By disliking automation in other seemingly simple linguistic contexts and allowing for outdating (of weeks perhaps), it seems that people much prefer the incorrect over the inaccurate. Which is fine as long as its sourced.
-SC
On 19 May 2010 03:59, Matthew Brown morven@gmail.com wrote:
I'm against the idea of automatically updating content for the reason that it's assuming the prediction in the article about a future event is true and actually happened as predicted.
Two examples of outdated statements, both spotted whilst browsing last night:
* "In September 2009, he will become chief executive of [the company]."
* "In early 2010, [the NASA facility] will begin preparing to support the new Constellation program."
In the first case, he did. In the second case, it didn't, because the program got cancelled. A self-updating text would be fine in the first half, but quite problematic in the latter.
Andrew Gray andrew.gray@dunelm.org.uk wrote:
In the first case, he did. In the second case, it didn't, because the program got cancelled. A self-updating text would be fine in the first half, but quite problematic in the latter.
Nice catches. I've been seeing them all the time too in my casual reading. I wouldn't have raised the issue otherwise.
Though I don't thing "quite problematic" is the right phrase, I do agree that in such cases we would need the flag to have a fallback mode. We've been discussing user messaging as a solution, where the switch tag whould include at least one editor's name attached, {{dateswitch|is going to fly|flew|DATE|~~~}}, and the bot would message the editor on the date to check up on it.
-SC