One of my colleagues sent round a link to a BBC news report from 2008 in the week. Chuckles aside, it’s the kind of thing that time and distance from the issue – geographical and emotional – renders insignificant in the grand scheme of things. It’s also the kind of thing that immediacy and proximity to the issue inflate into a big deal requiring things to be done to make sure this never happens again.

Often that’ll be some kind of instruction to immediately put more checking in place. Knee-jerk process is frequently fragile or deficient process and extra manual steps to secure an already manual process can easily have the opposite effect. It might be better to wait to see whether this is a one off or part of a pattern. It’s likely to be productive to spend a little time thinking about the whole, and alternatives, rather than simply trying to fix up the end.

Questioning a workflow, the need for the workflow, the implementation of the workflow and so on are tasks that a tester should be suited to. Here’s a few things to kick off, feel free to add more in the comments.

  • A standard out-of-office email in both Welsh and English would probably have prevented this specific case and would be a one-off cost.
  • Perhaps responses to requests for translation require pro-forma so that deviations can be recognised even if the content isn’t understood.
  • Perhaps all responses should automatically include translations (of both versions) by some tool such as Google Translate.
  • Perhaps there’s a more general problem with status of requests for translation. Perhaps a ticketing system would be appropriate. It’d also permit visibility of the number of requests, kinds of requests, time to satisfy, number of reopened issues and so on.
  • Other kinds of publishers will generally proof-read material, why are signs any different?
  • Why is this sign so wordy in the first place?
  • What was the process for the production of the sign? Was it followed? If not, why not?
  • Maybe certain positions need bilingual employees.
  • Maybe there’s other ways to have bilingual signage. Perhaps separate signs produced by a separate process using Welsh-speaking staff. Perhaps smart signs that alternate between showing English and Welsh, calling up translations from a remote server on the fly.
  • Maybe having less bilingual signage would be possible. The audience for this one is pretty specialised and there’s probably no-one who would be relying on the Welsh alone.
  • Maybe there’s no need to have any bilingual signage. According to Wikipedia, reporting UK census figures, 73% of the population of Wales don’t understand Welsh at all. 
  • Maybe Welsh should be the primary language with the English being translated from that. More people can proof the English translation.
  • Maybe more use could be made of existing or new language-neutral symbols.
  • Maybe there could be stock sets of dual-language signs that do require written language.
  • Maybe we don’t need so many signs at all.
  • How do we decide that a sign is required?

Be sure that the cost of any analysis is proportionate for the likelihood and severity of a recurrence of the problem. And don’t forget to fix the issue too: Swansea council made a new sign with a more accurate translation.