Google Groups has survived 22 years. How many self-hosted Discourse forums can say the same? If you're working on a new, small side-project like a new formal evaluator, are you going to want to spend your limited administrative bandwidth onboarding new contributors, responding to feature suggestions, talking to users, etc, or are you going to want to spend it trying to wrangle self-hosted forum administration and figure out email delivery? the article has a one-line tossed off aside at the end to say "Certainly self-hosted FOSS communities can die, but these are functions of community activity itself rather than the service they’re hosted on", but without considering that the absolute most crucial time for community resources to survive is through periods of lackluster or nonexistent community involvement. If you have a Google Group full of, say, formal methods programming experts, then it can spend 5 years fallow and still be chock-full of absolutely vital historical resources and even spring back into life as people start using it again for discussion. That sort of longevity just can't exist if someone stopped paying the hosting bills for their self-hosted TLA discourse forum 3 years in.
I co-admin a large group that had to find a new home when it was killed off. We went to a third party (groups.io) that accepts money in exchange for services.
Will that have longevity? No idea, but it's at least preferable to trusting [insert big tech co]'s benevolence in keeping a service (that they seemingly don't care about) running.
Money is but one piece to ensure longevity. How many people even work at groups.io? What's their business continuity plans in case the founder or someone critical to the team dies, either due to old age or some sort of tragedy? Or becomes tired of running it?
I'm not convinced it's preferable - people get old and tired of running a thing, or undergo some big life event that changes their priorities, and it's so tiresome to find a successor. Groups.io doesn't have the kind of transparency for me to trust in its longevity. (I am in a group or two hosted there though.)
Meanwhile, large tech can keep something running well past the initial creator's interest as waned. Sergey Brin and Larry Page left Google 20 years after they founded it, but I doubt they've had the same passion for new algorithms to improve Internet search for far longer. Despite that, Google is going to be around for a long time. Can you say that about groups.io? I don't even know who the CEO is or where the company is based.
In the case of google groups, the important thing to realize is that Google is from an older era of the Internet, when people used mailing lists to keep in touch and wrote long form posts rather than pictures on Instagram. So dropping support for mailing lists simply would not work with the current company culture, and something deep ingrained into a company's culture is difficult to change, especially as new employess get steeped into that culture.
I'm not going to deny the idea that Google Groups it's been "left to die", but I'd expect to see a facelift and a upgrade to Google Groups long before it dies, simply because it's an important dogfood product.
Groups.io was founded by Mark Fletcher[0], who in '97 started the mailing list service that Yahoo! eventually bought and turned into their Groups product. They're about 10 years in and he posts weekly changelogs and updates[1]. Nonetheless, they're small and one might guess there's a high bus factor here. (Sorry Mark!)
I think it's a toss up either way. I hadn't seen talk of Google Groups being internally significant until this thread, but they've burned enough similar products that it was barely a consideration for me.
> They're about 10 years in and he posts weekly changelogs and updates[1].
This isn't quite the brag you think it is.
10 years in, a product like this should be mostly fleshed out. On the linked page, I'm seeing design as well as technical changes mentioned. Also, they apparently only started working on translations recently.
Unfortunately, nothing lasts forever. Once you accept that, your perspective on 10 years may change. It's a long time. I challenge anyone to find discussions of this type from 100 years ago. Even if it was recorded (unlikely), did it survive after all of the participants died (unlikely)? Yes, we have 100 year old newspapers, but 100 year old discussions about underwater knitting.
Scientific journals were basically the mailing lists of their time, especially when they first started. You can easily find journals from 100s of years ago.
In theory, a good _independent_ product with an active user base and a profitable model has a better than average chance of surviving (bus factor aside).
Should the creator want to move on, someone should find the revenue stream attractive for some (even small) amount of money.
I’d bet on baseline profitability to keep something alive over an outsized dream of future growth or potential synergies almost every time.
I'm a normal enough person and so I can get good at most things. But, it'll take a few tries to get good at it. Riding a bike, swimming, SATs, etc. The first time I rode a bike was a disaster. I hurt myself and fell off. But I got back on and built up my bike balancing skills and, well, now I can ride a bike without falling off.
Soft skills, like managing people, are just like that. Another skillset to build, with practice, just like in an RPG. But one thing that I've never had to do before is find a successor to own my business. It's just not a muscle I've ever used. Nor, I bet, have many other people. So when Mark Fletcher, who seems like he's held this since before the Yahoo times, decides he's had enough and wants to retire, is he going to have a well practiced "find a successor" muscle? or is he going to fall off the bike, disastrously? Because the other thing about riding a bike is that you can read all the books, listen to all the podcasts about it, but that all pales in comparison to actually riding a bike and practicing.
You know who does have experience with replacing people like they're cogs in a machine though and nominating successor? Giant machines aka Google in this context.
So to each their own. I'm going to bet on dogfood keeping the dog alive until the owner decides it doesn't love the dog called Google Groups any more. I'm still very sad about Google Reader being put out to pasture and more recently Stadia, but lovable the German Shepard of a dog that is Search won't stop getting love from it's owner Alphabet and neither will Keep.
My problem is that the bus factor can't be put aside when it's such a small number, and that pile of "should"s add up to just as much, if not more, uncertainty, compared to Google letting groups living on in a half zombified state essentially forever. At the very least, you have to admit that once you put some thought into it, it's not as cut and dried as "boo hoo Google killed off Reader in 2011 and I'm still salty about it" (which, tbc, I am.)
>I co-admin a large group that had to find a new home when it was killed off. We went to a third party (groups.io) that accepts money in exchange for services.
Irony / wheel coming full circle, sort of:
Follow the links about M&A and founders in the links below.
Anyway, I knew about this history because I had a Yahoo! Groups account before it was that, i.e. was either eGroups or OneList, had been following Mark Fletcher, and had a groups.io account too, and had read his blog from long back.
I don't think it is necessarily a "problem" that it can't be perf-farmed anymore though. Google groups isn't perfect but it does a lot of things well (the bugs mentioned in TFA not withstanding), is used extensively internally at Google, and doesn't necessarily need a lot of new features. There are a ton of internal products at Google that are also staffed by one or two people in maintenance mode and it's not a problem.
As an example, I recently checked the internal symbolization service at Google (that handles symbolization for C++ executables, which is in turn used for generating performance profiles and symbolizing things like stack traces generated by core dumps) and the team maintaining it was two people. And those people aren't solely maintaining the symbolization service, they're compiler engineers who spend most of their time working on other projects and spend a small amount of their time maintaining the symbolization service as issues come up or bugs are reported. But this isn't a problem. The symbolization service works and is basically feature complete. There's no reason to have a team of ten people actively working on it.
> the team maintaining it was two people. [...] But this isn't a problem. The symbolization service works and is basically feature complete.
As long as it works, it is no problem. But given any larger change somewhere in the stack and management not being aware that these two folks have that side project, maybe they even left, and there will be problems.
Either a project has dedicated staff or it doesn't exist and just survives.
I don't know how critical groups internally are, but I assume with the general decline of e-mail compared to chat, video calls, etc. (insert joke about googles chat tools if you like ...) I would assume that this will be home less of relevance. Over time as well and drop more and more off.
Correct, and a good chunk of layoffs were people who never got promoted because they were tending the garden. So it’s actually worse than that, anybody who tried to maintain something is marked for deletion by Ruth’s fire-a-tron
That’s an interesting note, given that the Tie/X-wing fights were heavily inspired by WW2 dogfights, right? Can you imagine something that draws on the Gulf War similarly? It just seems wrong, too proximate to pull out purely aesthetic themes.
WW2 had an immense importance in the minds of people who lived it. Gulf War... who even cares about it, except for the ones involved? Many people don't even remember that.
Which of the gulfwars are you even talking about? - Even Korea and Vietnam, which were big and had quite some cultural impact, are more and more forgotten.
Although during the 80s and until Desert Storm, the same term was applied to the Iran/Iraq war - which is another mostly forgotten thing despite the huge human cost.
(I realised after I typed the above that the wikipedia disambiguation page also links to the Iran/Iraq war anyway)
Not the person who used the term, but I kind of wish I had been. It's excellent. At FAANG-ish companies where "impact" is the key to promotion, there's an inevitable tendency for engineers to prefer working on things that are highly visible like starting new projects or adding new features, vs. things that are highly useful like stability, testing, or code quality. Most internally promoted E6s or above got that way by initiating multiple projects, then dumping them on others when they've served their perf-review-enhancing purpose. "Perf farming" is as good a term as I've seen for it.
When you get promoted, they tend to move you to different projects (that need different resources at founding time than the operations folks that have experience scaling up)
I assume it's gaming the Google performance review system in order for promotion. That's why you see new Google products/features launched, then left to wither in the sun - ongoing maintenance/support doesn't do anything for your career.
I mean, it seemed objectively great when I heard about it for the first time, but then actually using it feels so icky that I’m deliberately looking for alternatives.
Google Groups basically destroyed all Usenet competitors and now is going belly up itself, likely taking down a large portion of computing history with itself.
Usenet was generally pretty good at sticking to binaries for the warez and copyright content. Hosting the text-only groups was not that intensive, especially by today's standards. I think the worst problem was sites that just didn't care... I spent a day reporting spam posts to the admins of origin servers, and they just didn't care at all.
Today, it's far less of an issue.. but if Usenet got popular again, the spammers would get back into it. I can only imagine how much spam gets removed/filtered on this site.
The Straight Dope has threads on their forum (which at some point switched to Discourse) from 1999. So given that you consistently have access to your own data I don't see why a dedicated webmaster couldn't keep a forum active for a significantly longer time than you can trust Google to keep any product alive.
Is it possible to export data from Google Groups? If not then longevity isn't the right question, they will not keep it active forever.
Not only that, but people continue to evolve the way they have (no pun intended) discourse. Email used to be the thing professionals and even friends out of reach did to message each other. I now use email effectively only for delivery confirmations and log in verifications. I haven't had a personal email in over a year I'd guess. People text, they whatsapp, they messenger, they telegram, but no one emails. The same with groups. A forum I have been on for decades has recently started a poll with the admins and power users of "Should we just close the forum since all the good stuff is on discord anyway".
Look at Slack's organic growth - it was great, people just started it and basically forced adoption on companies. Not so much today, because things evolved again. Discord is doing the same - people just start one, doesn't have to be official, then ultimately, it becomes official.
But you know, if your FOSS community can only survive with one specific tool, is it really a community?
> Email used to be the thing professionals and even friends out of reach did to message each other. I now use email effectively only for delivery confirmations and log in verifications.
Be careful with extrapolating from you personal experience. Pretty much all of my professional communication happens over e-mail. (I wouldn't even know what to use instead.) I am sure that is still true for many people.
I still visit some web forums that has been online since the 90's. At least two still use the same ancient software that they have used since ~1995. One of those is shut down a few hours every night (US time) for "database maintenance". I find that more adorable than annoying.
Used to hate web-forums. Usability is so bad compared to USENET. But now they are goldmines of probably unique content, and far better than Discord or Facebook Groups.
but it seems you only get 1 year of updates for the $160, and then you have to pay the fee again, so even if its lifetime use you will probably want to continue to pay for security or whatever else updates.
Honestly usually a forum dies because the community itself died first and no one cares about the thing anymore.
Google Groups may have lasted this long but as far as I’m concerned, it already died long ago and there was no reason to even go to it anymore.
There are still forums that out there that have survived two decades or more because their community is still alive. And keeping a community alive is not a technical matter.
Arstechnica forums have been going on since 98 or so. But they had that Conde Nast money early on.
What drives me bonkers is when newspapers just wipe out their historical comments sections. Like… archiving people’s thoughts is supposed to be what you do!
Right! I remember clearly when I registered on GTAF too. I'd discovered a bug/trick that allowed you to store a helicopter in a garage (I want to say in GTA 3?) and was itching to tell someone about it.
Of course it was already a known thing on the internet, but that was the start of my online career. From there I ended up on IRC and in there I wanted to make a bot, so I learned mIRCScript.
Spin on a couple years and I was helping run the IRC network, which got me into sysadmin stuff. From that rattling off how email works got me my first real tech job. "How in-depth do you want me to go?" "Uhh.. as far as you can"
I got the job offer before I'd finished with the DNS lookups haha.
> or are you going to want to spend it trying to wrangle self-hosted forum administration and figure out email delivery?
Nah, I'd just use Discord. Code rendering: check. With different syntax highlighting per language even. Discussion: check. Forums: check. Photo galleries: check. Email notifications: check. Free: check.
