hosting wiki and trac

104 views
Skip to first unread message

William Stein

unread,
Jun 7, 2016, 8:28:07 AM6/7/16
to sage-devel
Hi,

I am being pressured by various people to host trac and wiki in Paris
henceforth.

Technically, "the cloud" is now a far superior platform for hosting as
compared to "some guy with some computers at a specific university".
For starter, it is easy to make the compute available in any of
numerous geographic regions. Snapshots are straightforward. Access
to infrastructure is very clearly defined. The list of advantages is
big and getting bigger. I am fine with paying the hosting to have
the servers on Google Cloud Platform, and giving full access to those
projects to several admins.

William

--
William (http://wstein.org)

Volker Braun

unread,
Jun 7, 2016, 9:42:43 AM6/7/16
to sage-devel
+1 to hosting at at Tier 1 cloud provider

Also, if you have any arguments pro or con then send it to the mailinglist and not to William's private email. 

William Stein

unread,
Jun 7, 2016, 9:54:53 AM6/7/16
to sage-devel
On Tue, Jun 7, 2016 at 9:42 AM, Volker Braun <vbrau...@gmail.com> wrote:
> +1 to hosting at at Tier 1 cloud provider
>
> Also, if you have any arguments pro or con then send it to the mailinglist
> and not to William's private email.

+1 -- SageMath should be as open a project as possible. See, e.g.,
"Avoid Private Discussions" in
http://producingoss.com/en/producingoss.html for excellent arguments
for why.

Update regarding UW: the AC is back, and the servers are still off.
They will be turned on as soon as the sysadmin gets to work (it's
early there).

(I also work on SageMathCloud and we do have private discussions --
but these often involve specifics about users and their data, which we
must keep confidential. All dev work is still public on github.)


--
William (http://wstein.org)

Ralf Stephan

unread,
Jun 7, 2016, 11:25:33 AM6/7/16
to sage-devel
+1 for Google, I'm not a data-paranoiac, these are open data,
and better (ob)served by Google than the French gov (or any other nation).

Erik Bray

unread,
Jun 7, 2016, 11:34:51 AM6/7/16
to sage-...@googlegroups.com
On Tue, Jun 7, 2016 at 9:54 AM, William Stein <wst...@gmail.com> wrote:
> On Tue, Jun 7, 2016 at 9:42 AM, Volker Braun <vbrau...@gmail.com> wrote:
>> +1 to hosting at at Tier 1 cloud provider
>>
>> Also, if you have any arguments pro or con then send it to the mailinglist
>> and not to William's private email.
>
> +1 -- SageMath should be as open a project as possible. See, e.g.,
> "Avoid Private Discussions" in
> http://producingoss.com/en/producingoss.html for excellent arguments
> for why.

Sometimes there re good reason for private discussions. Think voting,
as an analogy. But final decisions should be made out in the open.

kcrisman

unread,
Jun 7, 2016, 1:36:26 PM6/7/16
to sage-devel

 
>> +1 to hosting at at Tier 1 cloud provider
>>

Just as a point of information, are there any other "Tier 1 cloud providers" who might be alternatives (usable for Sage development) for anyone who has some philosophical objection to one particular such provider?

Also, not to beat a long-dead horse, this is the kind of cost that a "real" Sage foundation would be able to provide funds toward.   I admit I'm uncomfortable having SMC revenues dedicated to this (I mean as the only source of hosting costs, not as a portion) but I don't know if the UW account is a feasible long-term solution for this.  That said, pass the hat!

Dima Pasechnik

unread,
Jun 7, 2016, 2:45:57 PM6/7/16
to sage-devel


On Tuesday, June 7, 2016 at 6:36:26 PM UTC+1, kcrisman wrote:

 
>> +1 to hosting at at Tier 1 cloud provider
>>

Just as a point of information, are there any other "Tier 1 cloud providers" who might be alternatives (usable for Sage development) for anyone who has some philosophical objection to one particular such provider?

the usual suspects: Amazon (AWS, by far the biggest), Microsoft (Azure), Oracle, IBM...

Erik Bray

unread,
Jun 7, 2016, 2:48:14 PM6/7/16
to sage-devel

This is a good question. (Probably worth a new topic.)  Is there any particular objection about approaching NumFOCUS about this sort of thing?

Reply all
Reply to author
Forward
0 new messages