[yoshimi] Re: Possibly bad news - a bit of history
- From: Ichthyostega <prg@xxxxxxxxxxxxxxx>
- To: yoshimi@xxxxxxxxxxxxx
- Date: Sun, 3 Jun 2018 17:57:38 +0200
On 03.06.2018 16:02, Will Godfrey wrote:
Originally Cal hosted Yoshimi on his own personal server, but (about 6
months) later transfered it to Sourceforge, as he said at the time "With some
reluctance". I suspect his reasoning was that with Yoshimi having quickly
obtained considerable traction and support he didn't want it to be easily lost
- well knowing his time was short.
However, when we lost him, Jeremy had to put in a fair bit of effort to get
Sourceforge to give him admin rights. It is fortunate for us that he did.
With this as a background, when I got involved I fairly soon gave several
people admin rights as further insurance.
When we started to have that series of outages at Sourceforge I then opened
GitHub mirror. The GitHub people to agreed to Yoshimi being a public open
source repository. This does two things. There are no fees, and it isn't
attached to me personally. Once again, I gave the same few people admin
Thanks for filling those bits of background story. I see your points.
However, maybe we should distinguish the various parts and services the
project needs to survive. Because it is precisely the main tactics of "THEM"
(be it google, apple or microsoft) to make people perceive their service as
an intangible unit, and suggest that people consider that to be without
An open source project needs
(1) a public source code repository
(2) a website that can be found by search engine. With a mission statement,
current links, some introductory instructions how to get and run the
(3) some kind of forum, mailinglist or social media group
(4) a bugtracker
(5) (optional) a download site with sufficient bandwidth to serve pre-built
executables plus an infrastructure to create/build those executables.
The numbering given above reflects the order of importance, as I see it.
Each of those has a quite different profile, both in therms of possible
vendor lock-in and with regards to the cost of a migration.
at 1) Git is a distributed system. The more people clone the repo, the
better the situation becomes. However, this depends on having active
code contributors. If only a single person does actual coding, then
a sole master repository can be a single point of failure.
at 2) This can be located anywhere and does not cost so much bandwidth.
But a migration is typically quite painful, since most offerings
create some kind of vendor lock-in. What Github provides here with
its Wiki pages is really quite valuable, since they reside by themselves
in a Git repo and use standard markup languages. Which means no vendor
at 3) these produce the highest degree of vendor lock-in. Plus there is a
network effect, since it is quite difficult to establish a location
which is actually frequented by people. Thus, especially the community
in the group/forum is that kind of asset the big companies are eager
to exploit. Thus: the only way to escape that kind of trap is to
run your own server (or get real donation / sponsoring from a
friendly umbrella organisation).
Yoshimi source code is available from either:
Our list archive is at: https://www.freelists.org/archive/yoshimi
To post, email to yoshimi@xxxxxxxxxxxxx
Other related posts: