Inert Detritus The Internet's dust bunnies

Indefinite Data Storage

No mat­ter how bright an idea it may seem at the time, set­ting “days of saved his­to­ry” to 500 in Camino is a Bad Idea.

It was nifty to have a huge his­to­ry of auto­com­plete URLs, and speed did­n’t seem to suf­fer much for it, but the fold­er orga­ni­za­tion with­in the His­to­ry view­er isn’t con­ducive to the idea: they have a sep­a­rate fold­er for each of the last sev­en days, and one last fold­er for “all old­er than 7 days”.

When that fold­er has over 20,000 entries, bad things result.

In gen­er­al, pro­grams, espe­cial­ly those designed for access the net, need to keep data reten­tion in mind. With appli­ca­tions like GMail (or setups like my site’s host­ing stor­age space being avail­able for IMAP-based email stor­age), users see no rea­son to lim­it the data that they retain over time. Even my Twit­ter feed, as far as I know, will be avail­able for all time, from start to end.

On the sur­face of things, “save every­thing I’ve ever accessed, cre­at­ed, or delet­ed” is a rule that makes sense, and ought be prop­er­ly sup­port­ed. Disk is cheap, CPU pow­er is cheap, and so indef­i­nite stor­age and access/search of all records should be an option. Some users want things delet­ed, and for good rea­son, but if I want to save every­thing, then I should be able to make that choice.