I've been avoiding this thread, but since I've been quoted I feel obliged to clarify my position. I apologize if anything I say has already been said, but I'm not going to take the time to read 8 pages of arguing.
Strictly from a developers perspective, an update checker is always a good thing. Practically all of their time associated with the mod is spent either reading bug reports or fixing them. I'm thoroughly impressed with someone like Reika who spends so much time doing that and still manages to add lots of new content. They have little incentive, aside from user pressure which I'll talk about below, to encourage people to play older versions of their mods. As Reika says, he gets lots of bug reports and complaints about old versions. Which, for obvious reasons, he ends up wasting lots of time trying to fix. So the update checker helps the dev.
Now from the users perspective. In this context there are three types of users. First, the average player. Whether playing an established modpack in single player, or playing on either a custom pack or an established pack on multiplayer, he has very little incentive to want to update a mod. In that first case, they are unlikely to want to spend the time or trouble to modify the established pack. Either they are a casual user who doesn't the expertise to find the correct download, replace the correct file, debug any potential conflicts and crashes, or simply doesn't want to spend the time. In the second and third cases, they don't have the control to update, since they don't run the server. In the fourth potential case, that of it being a custom modpack in singleplayer, they do control the modpack and slight incentive to update, more on the slight bit later.
Still from the users perspective, but now as a server owner. If they are running an established pack, they cannot manually update any mods. Doing so would require users to also manually update, which turns off the majority of users and defeats the point of having a unified pack. An update checker could provide incentive to update to a newer version of the pack, but again, more on that later. Second, if they are running their own custom pack, an update checker would potentially be helpful for them to see, if they were unaware of the update and were experiencing bugs that were fixed.
Lastly, from the pack makers perspective. Ideally, the pack maker knows as soon as an update is out. I can't speak for the Monster days, since I wasn't very involved in pack making back then, but now there's a constant stream of comments in our internal chats about mods updating. We know when a mod updates, what it fixes, what the potential conflicts are, oftentimes from the mod dev before it even comes out. As an aside, curseforge really helps with this. It lets you follow mods and receive an email when they update. It's an unobtrusive, opt-in system that's extremely useful for pack devs. Back to the topic, even though we know when an update comes out, we oftentimes can't update promptly. First, we have a release cycle. We release a non-recommended version, then usually a couple patches as we find bugs. Once the version has been stable for a week, we move it to recommended. The process from first release to recommended usually takes 2 weeks, but can run longer. That whole time, there are lots of people playing the previous recommended version, which itself was already 2 weeks out of date at the beginning, and is now 4 weeks out of date. This is a bit of an extreme example, FTB is large enough to have big enough bug trackers and user bases that we can justifiably wait that long, since not only do we know we can find most of the bugs, but the potential consequences of having to release a patch for a game breaking bug outside the normal release cycle. Like I said, a bit of an extreme example, most pack devs won't wait nearly that long. Potentially all during that time, mods would be informing users of an available update, which due to other factors, the pack devs cannot update to yet. This is not only potentially confusing, but can reflect unfairly on the pack author.
So we've established there are groups of people who could benefit from knowing an update is available, but there are also large groups of people who can have no benefit from knowing that, and in fact can be considerably annoyed by having it thrown in their faces all the time via an update checker. I have two more points to make, one of which is what a mod author can do to address the first group while not annoying the second, but first, a few comments on incentive to update.
Update checkers serve two purposes. The first, and obvious, is to let a user know an update is available. This can be useful both so the user knows to update, when applicable, but also to help prevent them from submitting bug reports for out of date versions. The checker's purpose should end there, but in a lot of cases it doesn't. If it continually prompts the user, it's providing incentive in and of itself. IE, annoying the user into submission, this is the second purpose. This is a universal problem in the software world, that companies have learned to avoid. It almost universally lends a bad reputation, for a non-obvious reason to most devs. That is,
most people who have your software installed don't actually use it. Bolded that part because it's really important. This might seem a bit presumptuous on my part, but it's nonetheless true. A big modpack has a lot of content. I could play for weeks and still not touch half the mods. In fact, most users end up doing something like that, though it would be a different half each time. I likely don't care in the slightest about updating a mod in that half. It's bugs don't affect me, new content doesn't really interest me. Prompting me to update it does nothing more than, as I said, annoying me into submission.
On to my last point now, what a mod author can do about it. If your intention is to annoy the user in to updating, I can't help you. If instead, you only want to make sure people know an update is available, I have a few suggestions. The most obvious is to go the path of other mods. Have the checker on by default, but allow someone a config option to permanently disable it. Since I can understand you being opposed to that, here are a few other ideas. Only display the update to moderators on a server and everyone in single player. Only display the message if the version is a certain amount of time out of date, ie, a few weeks. Only display the message if the user is on a non-recommended version.
For bug reports, I would highly recommend a robust tracking and reporting software. Something that aggregates bug reports serves two exceedingly useful functions. First, it gives each bug a single place for gathering information, you can mark new reports as duplicates and refer people to the main report, instead of relying on people reading many pages of a thread, often on topics unrelated to the bug. Second, it gives closure, instead of just saying a bug has been fixed, you can point people to the issue, giving all the details of the bug, when it was fixed, what version they need to update to to get the fix, etc.
Again, I apologize if any of this has already been talked about, or even implemented by Reika. I also tried to write it as a general comment on update checkers and intentionally not specifically directed at Reika.
Tl;DR I have yet to find an update checker that actually serves it's intended purpose of informing only relevant people of useful update information. Both learning about updates and addressing bug reports for out of date versions should be addressed in alternate ways.
Edit: Started writing this post about 3 hours ago, so I'll also address this question now.
Question: Of the two options, which is preferable?
- The update checker as currently agreed to: the closeable popup that can be disabled until the next major version with a command, which will have a config to make it so that only the pack maker can see it, until the pack is determined to have been abandoned (or at least has ceased being updated)
- Modpacks get no support at all. Period. End of story.
2, always 2. The burden of responsibility should ALWAYS be on the modpack author to address and replicate bugs before reporting them to the mod author. I would contend this also applies to your third party modifications policy, but that's another topic.