Aha!

[fivemack]

I know why the older proteins described in the PDB file format end each line with the PDB code of the compound and the number of the line in the file.

It's so, at the dawn of time when you carried around protein structures in boxes of punched cards, you could sort on the field at the end of the line and get the cards back into order if you dropped the box.

Comments

[bugshaw.livejournal.com]

♥ file format.

How many...

[seawasp]

... of your readers truly UNDERSTAND that the way I do, I wonder... (I learned COBOL on punchcards).

Re: How many...

[dd-b.livejournal.com]

COBOL on cards? Now that is really sick. Couldn't you define variable names that wouldn't fit on a single card?

[dd-b.livejournal.com]

I dunno if it's in fact true of PDB protein info; but it was certainly true of FORTRAN and assembly programs back when I started. They weren't required fields, but some of our decks actually had them filled in. And we did have an IBM card sorter with which we could have sorted them out of they'd gotten scrambled.

When loading a 3000-card source program into the card reader hopper, the chance of dropping a handful at least is non-trivial.

Dropping a box (2000 cards) could very easily pop the lid and spill the cards out. A *stack* of boxes dropped together made that even more likely. It could be a real mess.

[akicif.livejournal.com]

I'm mildly boggled, and now I feel really old - older even than "Bioinformatics for Dummies" made me feel.

There was a similar format for diffraction datasets - 3 sets of hklI (3I3,F10.6 I think)), then structure code and sequence number, but by the time I started doing crystallography, even the older of our two machines (the 3-circle with the wall-spanning PDP-8, not the cute 4-circle with the rackmounted PDP-11) had paper tape, so sequence numbers were already becoming redundant.

Of course, if I were to tell you my first structure was done with a Weissenberg camera, optical estimation of spot density, and Beevers-Lipson strips to calculate structure factors with, you'd probably think I was a lot older than I am.... But it did make a very nice basic undergrad practical (ammonium oxalate has a nice simple structure, is both monoclinic and centrosymmetric, so no weird special positions, and has a flat enough unit cell you can treat it as 2-d, so you only need a 0-layer Weissenberg.