On Thursday, 24 March 2016 at 09:39:34 UTC, Vladimir Panteleev
wrote:
On Thursday, 24 March 2016 at 08:41:18 UTC, Mike Parker wrote:
We recently had a suggestion her for a means of marking
threads as important or useful.
I really think this is entirely unnecessary.
"Sticky" threads on typical web forums are used to post things
such as FAQs or things people should read before posting.
I don't think he's talking about sticky placement, as much as
some kind of tagging of threads. Of course, that raises issues
of how you surface those tagged threads and whether people will
bother applying the tags.
not much left to object to. Ultimately, all serious open-source
software projects do their development on mailing lists. The
Linux kernel, Git, Gnome, KDE, LibreOffice, you name it. Can
you imagine someone telling Linus Torvalds with a straight face
that mailing lists are antiquated and it's time for him and his
gang to get on with the times?
Not only can I imagine it, I would say it to his face. Mailing
lists were incredibly outdated back when I first encountered them
decades ago, let alone today, which is why I have never used them.
The truth is that familiarity with mailing lists is simply
necessary for any serious software developer.
Not really, in fact, you can easily tell which dev teams are
ancient by the fact that they still use a mailing list as the
main form of communication.
Now, there are certainly benefits to SMTP/NNTP that centralized
forums don't have, no question, and a lot of web forum software
is ridiculously broken and even worse than a mailing list. But
you could do a lot better than a mailing list, it's just rarely
done. I though Apache Wave had some interesting ideas on
collaboration, though I never tried it, so I can't say if they
pulled it off.
Don't forget that forum.dlang.org has features that no other
forum software can offer, features many people depend on. That
includes its NNTP/email interoperability - one third of users
communicating on this group don't do it via the forum. (If you
think that one third is not too bad, don't forget that that
includes most of the core team.) The ratio will probably be
lower on "learn", but higher on the more technical groups.
The forum offers multiple view modes. Many people don't use the
default one, which mimics typical web forums. One view mode
I've added at Andrei's request, I think he will be unhappy to
see it go.
---snip-and-paste---
I am continuously collecting (constructive) feedback about the
forum. Last year I made an overhaul and implemented nearly all
feature requests. If you have specific requests for
improvement, please create a GitHub issue:
https://github.com/CyberShadow/DFeed/issues
All in all, I'm rather certain that as soon as an actual
serious proposal to replace forum.dlang.org with e.g. Discourse
appears, it will face just as much, if not more, vocal
disagreement. You can always create a poll or something if you
wish - out of curiosity, since as mentioned above, you'll have
a hard time convincing the people who are actually working on D
to switch.
As I understood what Mike originally wrote and he's now made
certain below, nobody is critizing DFeed for its features or
suggesting replacing it, only removing the lowest-common
denominator accomodation of email and newgroup readers.
The D forum also seems to be frequently lauded outside D's
community for its performance, and people seem to often present
in as an example of D's capabilities. It seems that any time
someone posts a link to forum.dlang.org, someone mentions its
unusually low response times.
Yes, I've seen that praise too, DFeed is a good showcase for D.
The "Save and preview" button was a great addition; I use it
often, particularly for long posts, and it largely obviates his
desire to edit a post whenever.
I'd like some sort of formatting language, like github has.
Can't you provide that option in the forum and send the resulting
HTML as text/html MIME attachments to SMTP and NNTP? I don't
know if NNTP supports MIME. Of course, some may complain about
HTML messages, but perhaps they can be handed some text
formatting fallback? Anyway, not a huge issue, but nice to have.
The current messaging status quo, where everyone gets an
undifferentiated stream of messages and then are forced to
manually scan the headings or run a keyword search on all the
contents, is incredibly outdated. However, advancing beyond that
will require some work, either to manually tag and vote on
posts/threads or write software that will at least automate
tagging, which is why it is rarely done.
But we need to move beyond this decades-old tech someday, as it's
wasting too much of our time.