[OSM-dev] [Tilesathome] ANNOUNCEMENT: T at H server will go away end of February

[Jeremy Adams]

On Wed, Feb 1, 2012 at 6:15 AM, Sebastian Spaeth <Sebastian at sspaeth.de> wrote:
> Hi all,
>
> I was just informed that the central ETH IT took an issue with the
> constant large bandwidth that the t at h server uses. In addition the
> server is four years old and getting old and senile...
>
> This means that end of February, the T at h server will be shut down and go
> away. Unless a replacement server (&admin) is being found, that will
> also likely imply that the T at H service will go away (together with the
> tiles web server) at that point.
>
> I have previously argued, why I believe that a t at h service is not that
> crucial anymore, although I still believe it is good that it was
> there. I am not sure if and what should replace t at h, but generally
> speaking there are so many ways now to customize and get rendered tiles,
> that t at h is a bit of a dinosaur. A fun one though :-)
>
> I would like to thank the ETH Zurich, and specifically the Institute for
> Cartography of Prof. Lorenz Hurni (http://wwww.karto.ethz.ch) that had been sponsoring the server and the bandwidth over
> the course of four years. They have been very generous, and never
> complained about the bandwidth/diskspace  we were hogging all the
> time. Finally a special thanks to Claudia Matthys, their IT sysadmin who
> helped to purchase, setup and administer the thing.
>
> Also thanks to the legions of renderers, client developers, and style
> tweakers, such as Bob Kare, Petschge, Dirk Lüder-Kreie, the ROMA and
> TRAPI developers, and the people involved in running the read-only
> mirror architecture. That was and is amazing work that diverse people
> have been putting together there.
>
> I am not that much into mapping any more, but I will hang around and am
> open to other fun projects.
>
> Sebastian
>

Boy, that's a shame.  :(  Always sad to see the end of a project, even
if it has served it's purpose.  t at h was the only way to get instant
renders after doing mapping work back in the day and really is/was an
extremely interesting collection of methods and technologies to make
that happen.  I've enjoyed all the time I've spent hacking away at it,
whether running clients or keeping TRAPI running along.

I will truly be sad to see it go.  (although it'll be nice to get my
20Mbit of outbound TRAPI traffic back ;))  I'd like to echo
Sebastian's statements above and say thanks to everyone who made it
happen.  There were many hands at work in developing and running the
t at h infrastructure before it because as stable as it's been lately and
they all desire a heartfelt thank you.

-Jeremy