and on Stack Overflow are quite ambiguous.
I'll cite computer-related terms from the Wikipedia disambiguation page for Clustering.

  • A result of cluster analysis. 1
  • An algorithm for cluster analysis, a method for statistical data analysis. 1
  • Cluster (computing), the technique of linking a many computers together to act like a single computer. 2
  • Data cluster, an allocation of contiguous storage in databases and file systems. 3
  • In hash tables, the mapping of keys to nearby slots. 3
  • The formation of clusters of linked nodes in a network, measured by the clustering coefficient. 1

There seem to be three popular meanings: various types related to cluster analysis (1), cluster-computing (2) and data storage/databases (3).

There are a number of related tags on Stack Overflow:

In my opinion these tags would be more useful on SO if we did actually not have "cluster" and "clustering" tags (which are inherently used for both), but typing "cluster" will instead bring up the suggestions [cluster-analysis] and [cluster-computing]. In particular, when the questions are mostly using these tags, they should show up in the beginning of the list.

What do you think?

  • 1
    1. Tag wikis where there are none. 2. Improve existing tag wikis. 3. What you wrote, +1. – yannis Jan 15 '12 at 12:12
  • I hadn't realised that you had brought this up on meta. Some of your retags are definitely not correct. You seem to have misfiled some type 3 ones. For example here and here have nothing to do with clustered computing. I'm not sure that database clustering is that relevant to the tag either but may be wrong on that. – Martin Smith Jan 15 '12 at 15:20
  • Good point, database clustering is the more appropriate tag there I guess. In particular, when you read "clustered computing" as "distributed number crunching", then it doesn't really contain "database server clustering". – Has QUIT--Anony-Mousse Jan 15 '12 at 15:34
  • @Anony-Mousse - Yes that's more what I would associate the term with. – Martin Smith Jan 15 '12 at 15:37
  • 2
    See also this feature request: meta.stackexchange.com/questions/119103, which should help with these kinds of disambiguations. – user102937 Jan 15 '12 at 18:06
  • @RobertHarvey yes. For example, "k-means" is a clear indication for "Type 1". While JBoss and Glassfish indicate a different meaning of "clustering". – Has QUIT--Anony-Mousse Jan 15 '12 at 21:31

already has a tag wiki describing it as , so we should probably merge into

Now that we have a tag, we can make and synonyms of .

We can make and synonyms of

And we can make a synonym of

And, of course, fixup all the tag wikis.

  • We should spend some thoughts on the exact names. As Martin brought up in the comments on the question, "cluster-computing" might to many not include the clustering of servers (database, web, failover, load-balancing), but true multi-core-computations. The better names we choose, the more likely users are to tag the questions right. – Has QUIT--Anony-Mousse Jan 15 '12 at 21:36
  • Perhaps add a [server-clustering] tag, with [clustered-server] and [database-cluster] synonyms? [dataclustering] needs to be fixed anyway; it's too close to this. I think [cluster-computing] is this. [database-cluster] could potentially stand on its own. – user102937 Jan 15 '12 at 21:39
  • I agree on dataclustering. The term "cluster-analysis" is much better. For this kind of "cluster-computing", tags such as "grid computing" or "supercomputing" might actually be more useful. But I figure it is okay. [server-clustering] sounds good to me, very clear. I'd keep database-cluster as an additional tag. – Has QUIT--Anony-Mousse Jan 16 '12 at 7:32

Not the answer you're looking for? Browse other questions tagged .