| Re: [tm-dev] TM download stats |
|
FYI, I also found out that all our downloads distributed via the Helios site had been counted differently … under /stats/releases/helios/ , numbers are:
I find these numbers suspicious, because ·
rse.core.pack (the core BUNDLE) is included in the rse.core FEATURE – yet we have fewer feature downloads (12692) than bundle (20510) ·
rse.dstore feature is included in rse runtime – so rse.core should be same as rse.dstore – yet numbers differ (14079 vs 12692) I also noticed that no statcounter had been in place for our bits in Helios SR1, so these numbers reflect the original Helios release only. Thanks, Martin -- Martin Oberhuber, SMTS / Product Architect – Development Tools,
Wind River direct +43.662.457915.85 fax +43.662.457915.6 From: tm-dev-bounces@xxxxxxxxxxx [mailto:tm-dev-bounces@xxxxxxxxxxx]
On Behalf Of Oberhuber, Martin TM committers, I have some interesting download statistics for TM: ·
RSE-core feature:
o
3.2 – ZIP runtime (4435) + ZIP SDK (687) + Updatesite (3908) =
9030
o
3.2.1 – ZIP runtime (4223) + ZIP SDK (662) + Updatesite (1920) =
6805 ·
RSE-SDK feature:
o
3.2 – ZIP (687) + Updatesite (4236) = 4923
o
3.2.1 – ZIP (662) + Updatesite (1749) = 2411 ·
DStore Server:
o
3.2 – Linux (527) + Windows (235) + UNIX (97) + Mac (40) =
899
o
3.2.1 – Linux (561) + Windows (244) + UNIX (129) + Mac (26) =
960 ·
Terminal feature:
o
3.2 – ZIP (241) + Updatesite (4371) = 4612
o
3.2.1 – ZIP (243) + Updatesite (1389) = 1623 ·
RSE-useractions:
o
3.2 – ZIP (63) + Updatesite (4289) =
o
3.2.1 – ZIP (97) + Updatesite (1324) = ·
Terminal-local:
o
3.2 – ZIP (59) + Updatesite (4096) =
o
3.2.1 – ZIP (69) + Updatesite (573) = ·
WinCE:
o
3.2 – ZIP (21) + Updatesite (3023) =
o
3.2.1 – ZIP (19) + Updatesite (409) = ·
Discovery:
o
3.2 – ZIP (43) + Updatesite (3541) =
o
3.2.1 – ZIP (40) + Updatesite (719) = ·
Examples:
o
3.2 – ZIP (83) + Updatesite (3478) =
o
3.2.1 – ZIP (113) + Updatesite (647) = ·
Tests:
o
3.2 – ZIP (52) + Updatesite (3389) =
o
3.2.1 – ZIP (32) + Updatesite (596) = Some interpretation of the data: ·
It’s very surprising that for less frequently “downloaded” packages (WinCE, Discovery, Examples, Tests) the “Update Site” numbers are still very high; in spite of the fact that those packages are very hard to find on the TM update
site, and not being published on the Helios site. This makes me think that Update Site Stats are likely not very reliable.
o
Either something’s wrong in the stats generation process;
o
Or some automated downloads occur on the update site.
o
I thus haven’t computed summary numbers for those packages since they are very likely wrong. ·
While ZIP download stats are roughly the same for 3.2 and 3.2.1, on the update site all stats are much lower for 3.2.1.
o
Again, this could indicate some “automated download agents” which had more time to ping the site for 3.2 than for 3.2.1 ·
A surprisingly high number of downloaders is interested in
dstore servers, and the distribution per Platform is interesting. ·
Discovery is still being downloaded more often than WinCE, although we have officially announced End-of-life for discovery. I’d be happy to hear more thoughts / interpretation. For those interested, the stats can be retrieved from
http://portal.eclipse.org in the “Tools for Committers” section. Use these substrings as filters: ·
3.2 ZIP Downloads: /dsdp/tm/downloads/drops/R-3.2- ·
3.2.1 ZIP Downloads: /dsdp/tm/downloads/drops/R-3.2.1-
·
Update Sites: /stats/dsdp/tm Excel sheet with numbers attached. Historical data for comparison is at
http://wiki.eclipse.org/TM/Download_Statistics Thanks, Martin -- Martin Oberhuber, SMTS / Product Architect – Development Tools,
Wind River direct +43.662.457915.85 fax +43.662.457915.6 |