Activate migration paths to Unix and Linux, should we?

Question

First of all, lets remember the golden rule of the migrations:

DO NOT MIGRATE CRAP

That said, why don't we open the migration paths from AU and UL? There has been 10 migrations to UL in the last 90 days, here the stats:

The migration to UL has been rejected lately, which has been normally in 0% but I don't really see where was the fail.

Some stats:

• In the last 30 days, there has been 200 question closed using the OT>not Ubuntu close reason, that is ~16% of all closed questions in that period.
• In the last 7 days, closed questions has been 44, or ~21% of all closed questions.

Relevant discussion on UL

Answer 1

EDIT 25-01-2020

Since posting this answer, I have spent a few years being a moderator on both Ask Ubuntu and Unix & Linux. I now have a much better idea of what users here suggest for migration and am completely, 100%, adamantly against opening a migration path. It's just not worth it, there are vanishingly few questions that would be worth migrating and what the community tends to think is worth migrating is usually really, really bad questions that would not be welcome on U&L.

---

I'd like to chime in here as a regular on U&L. First of all, I have to agree with Oli that our crap threshold is much lower on U&L. We tend to be far more aggressive with closing questions than AU and we seem to feel more strongly about having quality questions. That said, there are many Qs I've seen here that have been closed but would have been welcome on U&L. Some random examples:

• https://askubuntu.com/q/11405/85695
• https://askubuntu.com/q/24017/85695
• How can I manually change the default web browser?
• https://askubuntu.com/q/251850/85695
• https://askubuntu.com/q/269909/85695
• https://askubuntu.com/q/40910/85695

All of the above would have been perfectly fine on U&L. While there are extremely few cases where a question would be considered off-topic on U&L and on-topic here, there seem to be many cases of the inverse.

The only reason I am wary of this idea is, as suggested already by Oli, that AU members don't seem to be as aware of the SE model as other SE communities and I am afraid that U&L might be inundated by crappy questions if this migration path were opened for non-mods.

On the other hand, U&L is much smaller than AU and we can deal with a good deal more questions than we are currently receiving. If the AU community were to be diligent in not migrating crap so that U&L does not become a dumping ground for the, admittedly very many, bad questions that are posted here, I would be in favor of this proposal.

In summary, I think I can speak for U&L when I say that we would welcome your good questions that are off topic here but not at the price of a slew of bad ones.

---

As @derobert suggested in the comment below, it might be a good idea to just try this for a week and see how it goes. Unless it is very hard to implement, that should give us an idea of whether this is worth it or not.

Answer 2

I'm of the opinion migrations just don't work. I've said most of this before but I'll add in some Unix.SE-specific arguments

What is "crap" on Unix.SE and other sites?

I would say we have a higher tolerance for the crappier question than most sites. There will be lacking information (that we tease out in comments) or just failing to Google (how many bloody MergeList questions have you seen this week?!) but we help people nevertheless.

Now, if we go ahead with this, any future questions that dares even mention another distribution, whatever the current quality, people are going to jump on the migrate-to-Unix.SE-NOW button.

I think if Unix.SE want to petition to become a migration target from us, they need to look at the crap they'll be dealing with and realise that our users aren't always their users and they won't always adhere to their standards. That is going to be additional workload for reviewers.

We could easily be having this conversation again in a few months if/when Software Recommendations thinks it's ready to be a migration target. Their quality bar is so much higher again.

What's the follow-up rate on migrated questions?

I've been playing around with Data.SE and taking a look at questions migrated to Stack Overflow (just because they have the most in all directions). I didn't have time to build up a proper cascade count but a quick observation of the questions that were migrated:

• Many didn't have their users come over from the old site (denoted by no user account on SO)
• The vast majority weren't accepted
• Many had floating follow-up comment-questions showing the whole thing was just abandoned

We have enough of that without adding another way for people to lose track of their questions. I don't want to start shipping them in (or forcing them on other moderation/review teams).

If we push the person (rather than the question) they're more likely to end up in the right place, hopefully in the right way and they'll remain there.

Why does Unix.SE need to send us anything?

Theirs is a superset of our scope, if it's on-topic here, it's probably on-topic there. Only the most Ubuntu-only situations are ever going to be considered for migration and allowing for all other arguments past and present, their mods are very capable and experienced. If something needs migrating, somebody can flag it and they can migrate it.

Do our reviewers even know our scope? What happens if people disagree?

A running concern is that review here is not always consistent. Different people have different ideas and it's already true that people make decisions that others disagree with. We have reversals, mods stomp around and everybody's unhappy.

Given how many "Oh this would be better on x.SE because it's about programming" I still see littered around on questions that are perfectly on-topic with us here (the command-line tag has dozens), there could be a forced exodus of harder questions...

There is always going to be significant cross-over between the sites. A Bash/Ubuntu question is going to be on-topic here, Unix.SE, SU, SF, and SO... And probably more. I'm all for people advertising that there might be a better SE site but we pull our own even when it's not a question about Unity.

To put it frankly, I don't want our community scope lobotomised by a small group of like-minded reviewers who can't stand questions about Bash (or whatever). If you give people the option to migrate (without a feature like "counter votes") and that's what could happen. Even with a counter-vote, it could turn the entire thing into a tug of war over certain questions.

---

Migration is a dangerous tool that doesn't do what you think it does nearly as effectively as anybody would like. To work it needs to be user-centric and not just a process that reviewers churn posts through. It also needs safe-guards that don't currently exist.

The incidence where migration is needed is low enough that moderators can handle them manually.

Answer 3

Simply put - while it would be nice to have an all-in-one solution, they rarely work. Ubuntu is a Debian fork of Linux, with its own set of issues. Take, for example, the FFMPEG fight, which Ubuntu took sides in (rather than letting users decided between the two). Any who knows Ubuntu, knows how they handled the Unity desktop issue. These issues are specific to Ubuntu.