File Naming Limits
What is the current industry / server limitation on file names, specifically file name length? From the early days I remember this being rather limited but I am no sure now.
Personally I find it useful to have quite a bit of information in a file name. I know some shops deal completely with a number which is fine for some workflows but as soon as you leave that workflow you are stuck with a bunch of file names with little reference to what they actually contain. As an editor, I know this can be frustrating and time consuming once 'outside' of a CatDV workflow. Finding a workable balance can be useful in the future.
Are there any pitfalls for file name length? Obviously avoided strange characters and spaces. Our workflow will continue through to LTO5 (Cache A) so it has to stand up there too as well as server based use.
30 characters? 60 characters?
Any advice is greatly appreciated.
DOS used to have an 8.3 character limit and Mac OS 9 had a 31 character limit but modern Mac OS X and Windows let you have much longer names (up to 255 maybe?) but it's probably a good idea to keep to a lot less than that, say 60 or 80 characters or less, simply on practicality groups (filenames being listed in windows etc.). Also, there may well be shorter limits on certain file systems eg. memory cards, certain types of network file system, etc. And there are likely to be a lot of punctuation characters that may cause problems, most obviously : / but perhaps others besides. And don't rely on having two files with the same name that only differ in case.
It's a mixed bag. It goes from a GUID only approach (A432FDS.mov) to a full paragraph. I fall into the middle.
I'd love to hear if anyone has run into issues with long names. A larger issue I see all the time is bad characters. One of our recent docs we've developed for clients is our "bad characters" list to help get them clean up people's naming schemes.
What are you guys seeing out there as far as limits?
We recently did a several day engagement with a network on filenaming for them. We're about to deliver the schema but the basics are a prepended system that ID's the global info and then we allow the editors to use the end of the name for anything they like.
After my system sees "ID_6654_HX_AT" we don't care beyond that. their system includes share web docs with all of the codes and their keys in a web system and a full doc package so that new hires can follow along or at least be blamed when they don't. ;)
If this was just a MAM and we didn't have to deal with you dang editors/humans we'd have it easy. Machines don't care. ;) But in the end I also like having a place for plain text in a filename, but we also had to make sure the automated processes could run and that a human could also quickly see that a file didn't belong if a an ID_6657 ended up in a ID_6654 folder.
Obviously the specifics are confidential and the ID elements of each client get pretty specific but it's one example of something that every shop should consider whether you have a MAM installed or not.
Tip: If you have CatDV and a worker node, you can get a unique filename into your workflow just by having the Worker append or prepend the ClipID onto the filename so that you have unique names.
I've rambled long enough. Back to work.
bryson "at" northshoreautomation.com
I'll toss in a warning. You can easily run into a problem with the overall number of characters in the entire PATH gets too long. I've had it happen. I can't remember when, but there was a nice long string before ever getting to the file name. Error ended up being something like "excess of 255 characters, blah, blah, blah" and it took a bit to figure out it referred to the entire path.
I use old school naming conventions. I try to have a reel ID and a number. That's the tape name and it's relatively short and human readable, say SWA22 for Southwest Airlines 22. After that, well, depends on the system. Avid is going to make DNxHD files that are non-human readable except for the first few digits. However, you can name a clip whatever you like within the Avid bin structure. Final cut and Premiere are going to let you go all out. Still I keep it simple: MB15 cover 12.mov is going to be Missouri Baptist Hospital cover footage set 12. Interviews: RJ23 Mary Thomas interview 2 of 3.mov is Ranken Jordan Mary Thomas interview part 2 of 3, etc. And then lots of file-based systems have their own naming conventions that you change at great peril. But enter CatDV. The file name is simple and sweet but all the notes and metadata you add with CatDV is way better than sliced bread. I spent the day dropping time coded transcripts into the Verbatim Logger. Life is good when someone can skim the transcript and click on a timecode entry and bingo, the shot cues to it. Much more information than you could ever pack in a file name.
Oh, and the longer and more complex a file name is, the more perfect you must be in writing out the search string to find it. Learned that the hard way, too.
Non-linear: all the time and nothing but.
Great info as usual guys. The Worker node tip is excellent.
I used Long Path Tool software that simply worked for me for Long Path files.
This is a good thread, I'll keep it going too. This is not so much on limits but on the character sets that are "legal" for file names. Maybe I should break this out and post it as a new topic too?
On the subject of "good" cross platform file naming conventions:
There's what we are delivering to programming clients as a standard for file names.
An important thing to remember is that if you're storing data for long term use, it may get shared to another file system later (long after we've all retired or a software/computer company has gone out of business.) If you name things for Windows or Mac OS X only, you can get into trouble.
We recently ran into a nightmare where a company had named files using OS X rules, was using Windows automation and then their storage was a Linux system so that when we programmed using the APIs for all of these apps, we had to clean every name for all three OS's. Yes, a nightmare... and expensive.
They could have saved everyone months of work and thousands of dollars by simply using simpler file names.
So if your users are tempted to use a cute character in a file name, please stop them. lol The future depends on it. ;)
bryson "at" northshoreautomation.com