Diaspora has been a comedy of such errors for the decade it's existed
I think I had a Diaspora account in like 2011... Then they announced they were blowing away everyone's account to redesign the database or something... Never bothered with it again.
I wish I had been able to convince the guys to use a relational DB from the beginning instead of Mongo. Turns out a social network is full of relational data - but my buddies were too excited by new tech, and made a decision based on “cool”ness that hurt themselves in the long run.
However, it’s still impressive what four guys at the end of their college careers were able to build with no professional experience whatsoever. (Though it took me a full week to convince them to use a framework, Rails, instead of hand-rolling everything around a custom EventMachine loop... The loop never even worked, but it was “cool”. Give me boring and powerful any day.)
I think the biggest missed opportunity here was that an advisor of my own suggested to me that they really should leverage SMTP and build a better experience on top of that. An advanced mail client with social network-style features and presentation, and an obfuscation of the “pod” concept (do users really care which one you’re on?) could have been a smash hit and could have had fallback design to allow non-network-members to still be included as leaf modes in the Diaspora social graph. I couldn’t wrap my head around any of this at the time, so they never heard even a whisper of the idea. My bad!
> Though it took me a full week to convince them to use a framework, Rails, instead of hand-rolling everything around a custom EventMachine loop... The loop never even worked, but it was “cool”. Give me boring and powerful any day.
Funnily exactly that choice was what made me reject the request to help with development early on.
Not building on SMTP was a wise choice in hindsight, given that it turned into a pretty monopolized protocol since then - most people have a hard time to get their mails delivered from a private server nowadays.
That’s too bad. Their situation was that they were working in Pivotal’s office, surrounded by extremely high quality Rails devs who were happy to give free help and advice.
Like I said, the EM loop never worked, and instead of spending weeks banging their heads on it, Rails allowed them to move forward onto their actual problem, instead of having to invent everything themselves.
I guess I don’t know that much about SMTP still :)
Actually, I bet you could find people that would like the idea of a social media platform that wipes your history or disassociates your account every January 1st. Like a musical chairs, but the chairs never get reused.
Nice idea but it would only work if there were no ties at all between it and your real world identity, no connections that knew who you really were, because the usual method of harming someone via old social media posts uses screenshots, not links. The Internet has radically changed since back in the day, you cannot assume good faith of anyone.
Yeah, but that's getting to the very crux of communication: trust. If you're not sharing information it's pointless to call whatever it is 'social'. Anonymity is the form of least trust and you could argue the highest form of honesty. There's a very thick veil of soap scum on the Facebook ecosystem. Only the dumbest of the very tech illiterate are themselves.
Not to defend Diaspora, but it's not exactly like this stuff doesn't happen in real companies.
To pick one example out of a hat, when Twitter bough Smyte, they shut down the API 30 minutes after announcing the acquisition[0], leaving vendors like NPM with no warning. After getting hammered by large companies in the press who depended on their services, their response was, "we could have done better and are learning from this experience."
From a customer perspective, there's not really any practical difference between a company deleting your account because the founders were just given a ton of money, and a company deleting your account because the founders are worried about database performance.
A more accurate way to put it is that to put together a good service, back-end engineering is critical but is only part of the story. Good UX is also critical to success, and that comes with a focus on the human needs, not just the technicalities of moving data around.
What I meant was engineers made it for engineers. Dumping the database to change functionality is not something a successful user oriented system would do - but is something enigneers calling the shots would do.
It's something amateur engineers would do. Professionals should be able to handle database restructuring without losing data. Yes, it's harder, but... come on.
If the system is being that radically altered that you can't migrate data, just call it v2.whatever.com.... or point the old one to legacy.whatever.com, and redirect people with messaging.
Diaspora and Google+ and App.net really prove that funding and tech influencer press isn't enough to make a successful social network. Social networks are harder than they look.
If you look at https://joindiaspora.com/statistics, you'll actually notice that "Registrations" are "Closed". I guess that means that this specific "pod" is full and you need to find a different "pod"?
