[seqfan] Re: offsets of constants

Hagen von Eitzen hagen at von-eitzen.de
Sun Aug 23 19:28:18 CEST 2009


Charles Greathouse schrieb:
> I'm not entirely sure that the b-file format is best for constants.
> First, it's hard to read (by computers as well as people).  Why are we
> converting to a format that's not easy to use?  Pretty much every
> program out there natively reads and writes to one or both of
> 123456789
> or
> 1234\
> 5678\
> 9
>   
Indeed, such might be "c-files" instead of b-files (cf. there are 
already things like a-files for gappy ror "work-in-progress" versions of 
what might once become a fully grown b-file).
A proposed upload format might be to have a reasonable max line length 
(e.g. about 100 chars) and allow space for optional grouping.
Thus one prominent c-file might start

3.14159 26535 89793
23846 26433

I would suggest a simple script (e.g. perl) to create a c-file from an 
existing b-file and the %e line.
This can work even with a badly offset b-file and even with a wrong %O 
line - I guess the human readable form in the %e line is less error 
prone with respect to correct offset value than both %O line and b-file 
content.

Hagen




More information about the SeqFan mailing list