Timeline for What migration paths do we want on the site?
Current License: CC BY-SA 3.0
20 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 6, 2017 at 12:20 | comment | added | Emilio Pisanty | @peterh To be clear, my comment did not indicate that the query's (attempted) results were irrelevant to the issue at hand. By "meaningless" I mean that the (underdocumented) statistical methods in use are simply not powerful enough (i.e. intrinsically limited) to capture the data of actual interest with any kind of reliability. You're doing this with that query, pretty much, and the fact that you've caught the obvious correlations just means that your query can be replaced with someone listing the obvious correlations. | |
| Jul 6, 2017 at 12:14 | comment | added | peterh | @EmilioPisanty | There is also another systematic error source, which is most visible in the surprising similarity with the fitness SE: tags can have a very different meaning on the different sites, like "exercise" means some very different on the PSE and on the fitness SE. This is why this query, like any other heuristical calculation, can only orient votes, its results are not irrefutable. But, for example, I think the result that the PSE is more similar to the Chem SE as to the Biology SE, can be considered acceptable. | |
| Jul 6, 2017 at 12:12 | comment | added | peterh | @EmilioPisanty The - indirect - claim by mentioning the query in a post is that it can help the voters about their decisions, no more and no less. Yes, the special tagging customs of the MO cause a little deviation, which is partially fixed by two facts: 1) MO has synonyms to the usual tags (proof) 2) this query considers also tag synonyms. | |
| Jul 6, 2017 at 12:10 | comment | added | Emilio Pisanty | And on the overall question, you're not documenting any "minority opinion" ─ the only thing that's been conclusively documented is that this whole thing is entirely unnecessary, that the current mechanisms are working just fine, that some of your proposed changes would actively alienate other communities, and that the only changes that would even begin to make sense are currently out-of-bounds. If you insist on separating those two questions, then the OP as posed relies on the flawed premise that any of this is necessary or wanted by the community. | |
| Jul 6, 2017 at 12:03 | comment | added | Emilio Pisanty | @peterh I'm pretty amazed by the hubris implicit in the claim that some short SQL lines can suss out, say, whether a question tagged analyticity on PSE has related tags on MO (hint: it does, but your query won't find them). As a subtler example, something that would go under fa.functional-analysis on MO would happily be tagged mathematical-physics here; can your query detect those types of close relationships? I would think the requirement for context-aware NLP was obvious, but apparently not everybody got the memo. Sure, you can get some numbers out, but so can a monkey with a typewiter. | |
| Jul 6, 2017 at 10:56 | comment | added | peterh | @EmilioPisanty To your second sentence: Maybe you misunderstood the text of the initiative. It wasn't "Do we want new migration paths on the site?", instead it was "What migration paths do we want on the site?". Furthermore, even if the question score is negative, documenting or extending the minority opinion can be still useful, particularly if there is an intention or long-term plan to reverse this score. | |
| Jul 6, 2017 at 10:55 | comment | added | peterh | @EmilioPisanty About to your first sentence: this query calculates the correlation coefficient between the tag distributions, handling various problems (for example, inconsistencies in the tag synonyms) correctly. This is what it does, no lesser and no more. If you want to say that it is not related to the question topic, that it doesn't help anything about the voting preferences, I think it would require a stronger support as a statement that you want context-aware natural language processing. | |
| Jul 6, 2017 at 9:50 | comment | added | Emilio Pisanty | @peterh Unless your SQL can do context-aware natural language processing, I would think any such query is meaningless, but in any case I think the community opinion is pretty clear: this is a non-issue and it is not worth discussing it further. | |
| Jul 6, 2017 at 6:27 | comment | added | peterh | I wrote a query to measure the site distances by the correlation coefficients of the tag distributions. | |
| Jun 22, 2017 at 16:48 | comment | added | HDE 226868 | @peterh See the above. ^ | |
| Jun 22, 2017 at 16:47 | comment | added | HDE 226868 | Worldbuilding mod here. Two things: 1) Physics has migrated 12 questions to Worldbuilding ever, which is less than one every two months. Out of those, at least a third have either been rejected or deleted, meaning that you're sending us a good question less than four times per year. 2) As @KyleKanos pointed out, we're not a site for sending your crappy questions to. I have had to dissuade plenty of would-be migrations from Physics in the past because of this. There is no benefit in a migration path to Worldbuilding. At all. | |
| Jun 22, 2017 at 10:46 | comment | added | peterh | @EmilioPisanty I think you miss the topic. It is a survey, to get a clear picture, which migration pathes would be preferred by the community. I don't fight. In my opinion, the most constructive attitude to this survey is to simply voting up the migration path ideas what you prefer, and voting them down, what you don't prefer. If you have another, here doesn't posted migration target what you find useful, then you could extend the answers with it (ideally, as a CW post). | |
| Jun 22, 2017 at 10:39 | comment | added | Emilio Pisanty | @peterh You're fighting a non-problem. This is the kind of thing that would be worth fighting about if you had dozens of questions that were closed on this site that would be both migratable and valuable questions on the target sites. None of that is in evidence (despite the raw data being available for you to make that case) and until it is, this discussion is little more than wild hypotheticals and empty hyperbole on how useful the proposed changes would be. So: get your game up, and show that there is an actual problem. | |
| Jun 22, 2017 at 10:33 | comment | added | peterh | @KyleKanos Not all, but many of them. Typically, a significant part of the offtopic questions is coming from scifi idealists with only a lesser knowledge, if it results offtopic questions, then these questions may be well suitable for the WB. Note, the WB is a very inclusionist site, closure as offtopic is there rare. | |
| Jun 22, 2017 at 10:09 | comment | added | Kyle Kanos | @peterh: Note that WB ...is a question and answer site for writers/artists using science, geography and culture to construct imaginary worlds and settings. It is not at all for taking our non-mainstream questions and giving them reasonable answers. | |
| Jun 22, 2017 at 10:02 | comment | added | peterh | @ACuriousMind In my opinion, WB could be useful for many questions currently closed as "non-mainstream". But it is only a survey, maybe the community has a different opinion. | I didn't mention anything about bad questions, migrations are about the good, but offtopic questions. | |
| Jun 22, 2017 at 9:46 | comment | added | ACuriousMind Mod | @peterh You want to add worldbuilding to the list because there was a single question migrated there in the last 90 days, whose migration was rejected? Do you realize migration is not meant to be a way to shuffle bad questions around the network? | |
| Jun 21, 2017 at 19:19 | history | edited | Emilio Pisanty | CC BY-SA 3.0 |
added 397 characters in body
|
| Jun 21, 2017 at 19:07 | comment | added | peterh | 1) I think this should be a comment, with a link to the screenshot. 2) Having a migration path would surely increase the question migration probability. 3) Worldbuilding is a nice idea, I add to the list soon. | |
| Jun 21, 2017 at 19:04 | history | answered | Emilio Pisanty | CC BY-SA 3.0 |