|
Proposal: File Systems In my opinion, filesystems are tied closely to digital preservation. Preservation of data depends on having a robust filesystem supporting the data through decades of bit-rot(single-bit corruption), little maintenance, and constant additions of new information in the surrounding areas on disk. |
|||
|
|
|
If you were going to keep everything on the original disk, then filesystems may matter. If you are going to copy your Archival Information Packages to new [disk] hardware regularly, then it matters not so much. In our own setup, we store one copy of the AIPs on a Windows based server using NTFS, and another copy on a Linux based server using EXT4. We did upgrade the Linux storage from EXT3 some time ago. So, I'd say, in most circumstances, filesystems do not matter at all. The important thing is that the AIPS can be read by current systems. |
|||
|
|
|
There's an intersection, sure, but there's a lot to digital preservation that isn't related to filesystems and a lot do filesystems that isn't related to digital preservation. I don't see how they're related, as topics or audiences. I don't think the filesystems proposal is viable anyway. For design, there's Computer Science; for implementation, there's Stack Overflow; and for usage (which many of the example questions are about, even if it's not part of the proposed description) there are many IT sites (Server Fault for professionals, Super User for amateurs, and several platform-specific sites). |
|||||
|
|
I agree with the proposal. One remark though, you might want to mention filesystem in the title (metadata, sir). There is certainly an intersection and IMHO an important part of DP. One might want to use different filesystems for different purposes such as transport, preservation action copies, access, storage, etc. Things to consider: permissions, journaling, buffering, encryption (for non public archive records) and the like. |
||||
|
