FredBlogs The Ctrl+Alt+Del of ITD

Filenames and Windows

Filenames and Data Management
By phill

My Rules of Thumb for File Management:

1. Saving Files

* Always use ‘Save As’.
* Always save versions of important work or long pieces of work e.g.  save as paper1a.docx then later paper1b.docx
* Always save an email attachment that is useful in a sensible place straight away – if you are going to edit an email attachment do nothing until you have saved it somewhere safe – and work on the saved version.
* Use a synchronisation tool so you can access the latest version of a file wherever you are – iFolder is available at LSHTM.

2. Filenames

* Always use an 8.3 filename.
* Never use:  spaces, upper case letters, or punctuation except -(hyphen) and _ (underscore).
* Always name a file do not let Windows or software do it for you.
* Always turn on file extensions – i.e. the setting which reveals the extension part of the filename – Microsoft Systems leave this turned off by default.
* If you start a new file – name it and save it even before you have written anything.

3. Folders

* Never nest more than three folders deep.
* Always name them simply while still giving some indication of the contents.
* Follow similar conventions to those used for filenames i.e. avoid spaces, upper case characters, punctuation apart from spacers:  – and _
* Do not replicate information in the filename that is in the folder name e.g.
../acs-2010/acs-2010.xls

This collection of recommendations if followed will save you a lot of time and worry in the long run.  I will now outline a little of the reasoning behind this regime:

Using ‘Save As’ while more long winded makes you consciously aware of where the data is going.  ‘Save’ is quick but data is always lost when you are under the most pressure, or are at your most tired and vulnerable.  We often help people who have a very sketchy idea about where their data is going.  Despite our advice huge amounts of data are saved to ‘My Documents’ or the ‘Desktop’ – the locations that Microsoft bullies us into using.  Fine for a static standalone machine at home but not suited for use on a networked workplace system.

Few people seem aware that there are rules about what is allowable and possible in Filesystems.  Microsoft filesystems have a limit of 255 characters for a filename – that is including the entire folder pathway and the bits like c:\Documents and Settings\My Documents\blah blah.docx – 53 characters worth.  There is more to all this – have a look at :

http://msdn.microsoft.com/en-us/library/aa365247%28VS.85%29.aspx#naming_conventions

for a more technical explanation.

More useful information can be found here:

http://www.portfoliofaq.com/pfaq/FAQ00352.htm

The point is to take control of the data.  Simple names following a logical scheme makes data easy to organise, easy to archive, easy to manipulate on all available platforms e.g. Unix, Linux, Mac, Windows, and a multitude of more esoteric operating environments.

I want to show you an example of a file spath that has recently caused me problems:

‭\\‬Itddata5‭_‬server\ITDDATA5‭\‬TEMP\IDCVxxxx\FRI3007‭\‬ITD\DCV\HOME\IDCVxxxx\notes and scraps of work\very old sequence data\BY DATE\march‭ ‬2008‭\‬raw sequences\old sequences\LSHTM sequence analysis\sequence data\BY DATE‭\‬Nov‭ ‬2007‭ ‬sequencing\plate‭ ‬2‭\‬fasta seqs\LC_P2_M13F_A10_080_A19.seq

This was the path after the data had to be relocated in the system;  300 characters in total. Now when this folder full of 581 files was created its path would have been:

H:\notes and scraps of work\very old sequence data\BY DATE\march‭ ‬2008‭\‬raw sequences\old sequences\LSHTM sequence analysis\sequence data\BY DATE‭\‬Nov‭ ‬2007‭ ‬sequencing\plate‭ ‬2‭\‬fasta seqs\LC_P2_M13F_A10_080_A19.seq

226 Characters.

This caused problems with copying the data to the new location – the copy software failed – fortunately using a proper copy program instead of Window’s copy there was at least a log of the files that did not make it and why.

Technical stuff aside – it makes sense to aim for simplicity and elegance in all organisational tasks

By planning a system of data naming and data organisation you are taking responsibility for your work and taking control back from the operating system – tacky old Windows for example – and empowering yourself. Building an efficient and elegant data strategy gives you an inner mental map of the data and gives you an overview of the structure of the materials you are working with.