Allowlist, not whitelist. Blocklist, not blacklist. Goodbye, wtf. Microsoft scans Chromium code, lops off offensive words

Redmond suggests nuking 'profanity, geopolitical, diversity' terms from browser source

Cleaning the office

Microsoft's adoption of the Google-developed Chromium browser engine for Edge has resulted in a proposal to cleanse the open-source code of "potentially offensive terms."

Issue 981129 in the Chromium bug log lists a suggestion by Microsoft to “cleanup of potentially offensive terms in codebase” aims to rid the software blueprints of language such as whitelist (change to allowlist), blacklist (change to blocklist), “offensive terms using ‘wtf’ as protocol messages,” and other infelicities.

This bug report was raised by a Microsoft contributor, who stated: “We are just sharing a subset of what PoliCheck scanned for us,” Policheck being “a machine-learned model that another team manages that does context based scanning on hundreds of file formats.”

Googler Rick Byers, a Chromium engineer, gave the issue a cautious welcome, saying: "This sounds like a good strategy to me, thanks for doing this! We certainly have never intended for anything in the codebase to be potentially offensive, but I'm also not aware of anyone making an effort to find them all." He added:

I don't expect Chrome teams to necessarily make these bugs a priority (we haven't seen this pose a problem for us in practice as far as I know), but if cleaning this up is valuable for Microsoft (or any another Chromium contributor) then we should have no trouble getting the necessary code reviews (at least in the platform code). And yeah there are folks who look for GoodFirstBug and may want to pick up some easy commits.

Although changing comments or variable names in the source code is generally invisible to the user, this kind of revision can be problematic if it wrecks things like names in preferences and policies.

Judging by the relatively short list of issues the Chromium source code is already – for the most part – clean, as befits such a public body of code.

No swearing WTF?

What the @#$%&!? Microsoft bans nudity, swearing in Skype, emails, Office 365 docs

READ MORE

Opinions will vary on the value of such changes, though if Microsoft is serious about its diversity and inclusion policy, it is not surprising that language used in source code falls under these rules. Google also has a pro-diversity program, as you'd expect.

In May, Microsoft announced AI features in Word that, among other features, will emit “advice on more concise and inclusive language such as ‘police officer’ instead of ‘policeman.’"

The PoliCheck tool seems to have been around in some form for a long while. “Reading that Policheck file was … incredibly educational. I learned all sorts of exotic curse words I didn't know before,” remarked developer Andrew Gaspar, who described himself as “formerly Windows Core Engineering at Microsoft.” The discussion concerned the Windows 2000 source code leaked in 2004, said to contain many references to idiots and morons though “no racist or homophobic slurs.” ®

Sponsored: Technical Overview: Exasol Peek Under the Hood

SUBSCRIBE TO OUR WEEKLY TECH NEWSLETTER




Biting the hand that feeds IT © 1998–2019