Hi Denis,
If download stats deserve a higher SLA, perhaps engage with the
Committer Board Representatives for guidance.
Thanks, I'll investigate that.
Querying download stats like we do is quite passé.
Being passé isn't necessarily a bad thing (isn't the average Eclipse.org project mostly passé?).
Current query engine is extremely powerful and allows to get more useful input that any pre-generated download stats. Download stats can help decide whether to abandon specific versions, specific artifacts, can allow to get fine grain feedback on relatively small changes on one day after the other.
IMO, it's much more important to ensure the current capabilities of download stats are reliable than to build new visualizations/projections.
However, it seems to me that since last bug fix, download stats are reliable, aren't they? That's a genuine question, this was discussed on the ide-dev mailing-list and not all committers trust data the same way and don't have the same feeling of reliability. What are the limitations, possible glitches (potential part of bots, of builds, are user-agents checked...)? Is this "quality"/"reliability" made explicit somewhere?
Thanks in advance,