Hi,

There has been a string of unexpected obstacles on our path.

1) sg-int1 crashed several times, once during Ansel’s onsite visit,
after Basheer (ECS engineer)’s first visit and again shortly after
Basheer’s second visit.

2) We wasted time waiting for sg-int1 to be repaired. We had earlier
done quite a lot in sg-int1 and we wouldn’t want to repeat them in a new
box. In the end we gave up waiting for sg-int1. A mis-calculation that
cost us a few days.

2) browser access to the subversion system requires a special binary of
apache, with mod_dav and mod_dav_svn. We tried very hard to avoid a
source rebuild of apache and svn. This research took quite a long time.

3) The very first svn install (minus apache) was also rather tedious
owing to the required packages. Such a complicated procedure would be
hard for colleagues to follow. I was confident it could be simpler. The
research took some time.

4) The archive was bigger than expected, and took several hours to
transfer by ftp. We realized the data transfer rate was way below
network capacity, so we spent some time discussing and trying out other
ways to transfer the file. This effort took some additional time.

5) The checkout process took about half a day, which was not known when
we were planning the timeframe.

6) The requirement was not well-understood in the beginning.