4 ms·

Probably the best file naming convention ever (2014)

11 points by beauHD 5y ago

dcminter 5y ago
I sort of hate this. Everything in it is reasonable, but none of it should be reasonable. Except point 3 about using ISO-8601 date formats, which is merely righteous.
- m-p-3 5y ago
  Personally I prefer ISO8601, it's unambiguous and easy to sort in chronological order.
adamredwoods 5y ago
The naming of things is such a debatable topic, but such an important one. I've seen the name of programming functions with odd phrases, that might not be clear if someone wasn't a native English speaker. I've also incorporated naming structures for off-site jobs, using job numbers as a prefix, and then a 'readable' word afterwards. That seems to work. Lately, I've picked up making my git branches with name/service/subject/something-additional (for example). This seems to keep things organized instead of having to search through mountains of files. Even the naming of my family's photo files is still under debate. Doing just year/month gives no context, and usually we search by subject (camping at Mt Shasta, for example). I don't remember the specific date. Or what about first child walking? I don't remember that date either, but having a year/subject/month folder may help. And restructuring is a pain! Any method started is usually retained.
bradknowles 5y ago
In some cases, I think using a heap storage mechanism plus tags is a much better solution. Then things like tags for date and tags for location can become at least partly automatic. Coming up with a single canonical naming system where each thing can live in one and only one place is not a good solution in many cases, like photographs. In general, naming of things is the only universal “hard” problem that I know of. Each subject area has their own hard problems to deal with, but all subject areas have to deal with naming. And sometimes the best solution is to tag and not name at all.

4 ms·

Probably the best file naming convention ever (2014)

11 points by beauHD 5y ago

dcminter 5y ago
I sort of hate this. Everything in it is reasonable, but none of it should be reasonable. Except point 3 about using ISO-8601 date formats, which is merely righteous.
- m-p-3 5y ago
  Personally I prefer ISO8601, it's unambiguous and easy to sort in chronological order.
adamredwoods 5y ago
The naming of things is such a debatable topic, but such an important one. I've seen the name of programming functions with odd phrases, that might not be clear if someone wasn't a native English speaker. I've also incorporated naming structures for off-site jobs, using job numbers as a prefix, and then a 'readable' word afterwards. That seems to work. Lately, I've picked up making my git branches with name/service/subject/something-additional (for example). This seems to keep things organized instead of having to search through mountains of files. Even the naming of my family's photo files is still under debate. Doing just year/month gives no context, and usually we search by subject (camping at Mt Shasta, for example). I don't remember the specific date. Or what about first child walking? I don't remember that date either, but having a year/subject/month folder may help. And restructuring is a pain! Any method started is usually retained.
bradknowles 5y ago
In some cases, I think using a heap storage mechanism plus tags is a much better solution. Then things like tags for date and tags for location can become at least partly automatic. Coming up with a single canonical naming system where each thing can live in one and only one place is not a good solution in many cases, like photographs. In general, naming of things is the only universal “hard” problem that I know of. Each subject area has their own hard problems to deal with, but all subject areas have to deal with naming. And sometimes the best solution is to tag and not name at all.