One other option -- cite something announcing / reviewing the particular release. If you have to, even cite the perldelta for the version. I would assume the author for the perldelta would be the pumking for the release.
Update:
And another random thought (and why I mentioned citing a reference, rather than directly) -- there's been an issue in scientific fields regarding how to cite the actual data used in the research, and I saw a talk last week at CODATA by Sünje Dallmeier-Tiessen, where she discussed an journal for the publication of earth science data. They have a standard template for presenting important aspects of the data (where to get it, provenance, coverage, etc.), and have a peer-review process ... so that people can cite the article describing the data set, and standard bibliographic tracking will pick up the relationship.
I have no idea how often it's necessary to cite programing languages, but if it happens often enough, it might be worth people working together to create a similar journal for people to announce major updates to programming languages.