[Webkit-unassigned] [Bug 104858] "Running 1 DumpRenderTree over X shards" is not a helpful output

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=104858





--- Comment #4 from Dirk Pranke <dpranke at chromium.org>  2012-12-12 17:20:41 PST ---
jargon is often, but not always, bad. Sometimes inventing new terms to accurately and concisely describe concepts is the right thing to do (I am not necessarily making that claim here, mind you).

All of Maciej's suggestions in comment #2 have the downside of being vague and interpreted as one of many possible concepts. In particular, NRWT already uses "part" and "chunk" to refer to different subsets of the test as well. "shard" has the virtue of having fairly specific connotations, even if they are being applied slightly differently here than one might be used to seeing them.

As I mentioned over #irc, in the server clustering / cloud community, a shard is *not* necessarily a separate physical resource; shards are virtual concepts that get mapped onto physical resources, typically in an M:N manner (which may of course be 1:1, but the whole point of a shard as opposed to a parallel database is that you can move shards between physical resources to redistribute the load, which is exactly what we're doing here).

"batch" or "job" might be better terms that are at least closer in spirit to "shard" than something generic like "section". 

Unfortunately, we also use --batch-size in NRWT to mean something else. "suite" is also close, but not quiet right (since people might get confused when a suite doesn't refer exactly to a directory).

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.