Hacker News new | past | comments | ask | show | jobs | submit login

Usenet news.

1. You enter groups devoted to large topics (like "all company" or "software development" or maybe "Smith Project")

2. You see conversations delimited by subject lines, completely threaded, with full history.

3. You normally reply to a message with another message, quoting or not, and your message is threaded in.

4. The tools are good for showing you what you haven't read yet.

5. You can ignore a topic forever.

6. You can have filter highlight topics where your name or email address or a keyword is mentioned.

7. You can seamlessly fall back to email for private conversations.

8. There's an archive that goes back to the limit of storage space.

9. Usenet is, of course, easy to distribute and federate.




Yep. That was exactly my thought.

Everything old is new again. (And Usenet is even still there.)

I'm still convinced that Usenet gets ignored precisely because it is federated and mature - it is harder to extract nickels from than a centralized service; user-surveillance and other ad-related choke points are fragmented, it works already, there are already robust end-user filters.

(I also believe Twitter should have been an RFC, not a startup.)

For anyone who wants to build similar stuff for internal use - NNTP is probably not going to be the right choice unless you have a large number of sites. For a startup, IMAP has everything you need built-in for shared, threaded group messaging that looks a fair-bit like usenet (from the client's perspective). If you don't have strong ideas about UI or workflow, you can use an MTA to read and respond, and you're done. If you want your own UI, there are a huge number of IMAP libraries out there for just about any language you want to use.

"Old" tech is a goldmine.


> I also believe Twitter should have been an RFC, not a startup.

I have this reaction to _so many_ online services these days.


It would be fun to actually write the RFCs for those services. Besides, perhaps they could trigger open source and federated alternatives.


"Fun"? I incidentally starting writing a RFC-like spec and it's sooo tedious to define all the little details. (I don't mean that it's not useful. It helps me very much in putting my ideas on a more rigorous foundation, but damn is it tedious.)


Wow, I love all these features! You're not saying it:

10. Has open-source implementations,

11. Allows HTML formatting,

12. Has a hierarchy of topics, splitting up topics in a generic way,

13. Is blazingly fast,

14. AND allows referencing specific articles using an URI scheme: https://tools.ietf.org/html/rfc5538 ?


Blazingly again? I wish people stop using words only because it's cool to do so.


I'm a non-native speaker. I'm guessing you're saying the adjective is dated and reading it again it does sound like that. Still, i was trying to say 'faster than the usual fast'

Exorbitantly fast?

Snappy responsive?

Real-time? (Haha)


Usenet was not historically real-time because of several things:

1. There were a lot of servers.

2. The servers were not arranged in an optimal network.

3. The network was full of cross links, many of which made no topological sense.

4. Not every server carried all newsgroups.

5. There were a lot of users, who read a lot more than they write.

6. Binary messages (when carried) grew to huge sizes (for the time).

7. Network links were very slow by today's standards.

8. Disks were slow, small and expensive by today's standards.

9. Every ISP felt it had to provide free Usenet service, but few of them did it well.

With modern hardware, Usenet could be as close to realtime as you expect email to be -- dominated by people's attention and writing speed. And carried over TLS, of course.


There are internal message board systems in Microsoft shops (i.e. Sitrion on SharePoint) that struggle to achieve 2,3,4,5,6,7,9. Despite beefy servers and huge $$$, still slow and lame.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: