sequence trace files format problem

Mike Parlee mparlee at cts.com
Thu Sep 30 21:39:38 EST 1999


In the past I have handled this problem by serving the files from a Unix or
Linux box running Netatalk (http://www.umich.edu/~rsug/netatalk/).  With
netatalk, a unix volume can be mounted by macs on the network and the traces
can be dragged and dropped into the folder just as if it were another mac.
The resource information is stored in a shadow directory (AppleDouble) and
the files are remain useable both by the mac and by the unix box (or Windows
through Samba).

You do run into a problem if the files are moved around on the unix side
though.  If you are going to manipulate filenames or move the files under
unix you will want to configure netatalk to map Macintosh file types and
creators to a dos or unix file extension as follows.

In the AppleVolumes.system file add the line:

.abi    ABI1    ABI1

This will associate any file ending in .abi with a psuedo-resource fork
containing the Macintosh file type and creator for an ABI trace file.  Use a
script to append the .abi extension to any ABI files that you manipulate
under unix and a Mac connecting to the unix box will still be able to use
those traces.

I believe this can also be done in Windows NT server, but so far I haven't
been able to make it work.  If anyone knows how to map Mac file types and
creators to an extension in WinNT please let me know!

Mike Parlee
mparlee at cts.net

> Paula Burch wrote:

> > The problem is that trace files moved to the sequence repository
> > machine from the sequencing Mac can be used only by a Mac, not by
> > Unix, if tranferred as a Mac binary. If the trace files are moved as
> > 'raw data', they work great in the java trace viewing program, but
> > are utterly unusable on a Mac.
> >
> > How can I make 'raw data' trace files work on a Mac? I was told
> > something I didn't understand about a missing 'resource fork' that
> > the Mac needs. I generally stick to Unix and am mystified. Is
> > there any way to fake the resource fork, or any other way to make one
> > transfer format work for all three of the more popular computer
> > platforms, Mac, Unix, and PC?
> >
> > ________________________________________________________________________
> > Paula E. Burch, Ph.D.             Molecular Biology Computation Resource
> > Baylor College of Medicine, BCMM M220                voice: 713.798.6023
> > One Baylor Plaza, Houston, Texas 77030                 fax: 713.798.4279
> > http://mbcr.bcm.tmc.edu/pburch.html            email: pburch at bcm.tmc.edu





More information about the Bio-soft mailing list