Monday, February 15, 2010


There appears to be more than a bit of unworldly nonsense about how much work it is to respond to FOI(A) requests, Jules explains

Contrary to the naïve belief I read all over the internet, fulfilling an FOI can be much more work than just taking the data out of the drawer. From my own experience I can say that it can incredibly time-consuming to collect historic data. For a project at my previous job I needed to find as much interesting data as we had in our archives.

The only way to do so was literally digging in the archive and sieving out the data I could use. After I gave up looking any further, my best estimate being I managed to collect 70-80 % of all the data available. Before you go all crazy : It does not mean the other data is missing, it just means it would have been too great an effort to find them. Some of them are hidden in metres and metres of papers and reports, other are in an old database that has been out of order for years, etc.


The tools to beat back abusive FOI requests are simple. First, there should be a limit on how long a search takes before charging (full rate) those who made the request. Second, there should be a limit on how many hours per year any one person can be tied up by FOI requests.

John Mashey may remember this portable memory device



David B. Benson said...

Ah yes, 7 hole punched paper tape. I used to use a hand punch and specialized Scotch tape to literally patch my programs.

Those were the days...

Anonymous said...

Hell yes!!

I can't begin to explain the difficulties FOI requests involve.

I have spent many a day in old sheds looking over silverfish eaten, mouse dropping covered, worthless correspondence. My old bosses used to save every carbon copy of every letter they'd ever sent. The office had been moved 4 times and junk was left in storage buildings for years on end.

Inevitably I'd get a FOI for every correspondence ever written on one subject or another and knowing that if someone sued us and was able to produce a single record that I'd missed; I'd go through their old boxes looking for crap.

Eventually our agency developed a policy to toss old, worthless junk.

Anyways Scared Mouse says what the hey, it's a living.

John Mashey said...

Yes, we have some of that at the Computer History Museum.

Scene: I'm leading a small group of first-term Stanford students the accelerated computer science program ... i.e., these kids have written C++ for years, put together Linux PCs, etc.

I show them replica of original Hollerith machine (for census), then early IBM keypunch, and a few punch cards, noting that I used to punch whole decks of them.

Them:Why would you do that?

Me: Actually, that was the only way to write programs at that point.

Them: (from the group) What? Gasp! (muttering) Nobody could write programs that way. You have to be kidding. (People actually used this junk?)

Me: Wait till you see paper tape.

Rattus Norvegicus said...

The first machine I programmed was a Wang. It had an optical card reader. Writing a program involved coloring in the opcodes with a #2 pencil, colored in was a 1, not a 0.

dhogaza said...

Looks like PDP-8 binary format, to me. Each 6-bit doublet formed a 12 bit word, and the 7th bit indicated that the 12 bit word was to be interpreted as an address, with the subsequent doublets stored in consecutive addresses until the next address doublet's reached.

I counted the rows off from one row with the 7th bit punched to the next, and yes, it's an even number.

Unlike John, I was lucky enough to never have to punch cards manually, having grown up with the PDP-8. TTY and paper tape, later DECTape. As a freshman in college I got full computer center rights to our CDC 3300 and typed stuff into a very earlier video terminal. And then punch out a compressed binary deck version of the source for backup...

One thing the young'uns are missing is "1,000 things you can do with chad, most of them obnoxious". Especially chad from oiled paper tape.

Deech56 said...

Do the young 'uns even get "Do not fold, spindle or mutilate"?

Martin Vermeer said...

Yep, I once wrote a screen editor in Fortran to use on an old ADM-3A 'glass teletype' terminal, as it stood mostly unused while the few expensive vt100's we had were always booked...

Anonymous said...

Droped deck, shuffled cards, endless loop, end of career.

skanky said...

In Leaky Establishment by David Langford, one character plays space invaders on the teletype.

I've sometimes wondered if there's any basis, seeing as the rest of it has some pretty solid grounding in reality?

EliRabett said...

First time Eli remembers playing space invaders was in 68 on a new DEC computer the high energy group got with a video display (not a terminal) so it may have used a teletype.

One of the Bunny's friends was still taking data in 2002 on a PDP-8 that you had to enter the boot sequence from the front panel switches. His students were happy when that died (it may have been euthenasia).

John Mashey said...

But further back, the COmputer History Museum has restored to workign order:

IBM 1620
IBM 1401

The first two sit in a room refubished to look like the DEC "Mill" in Maynard (wooden floor), the third in a raised-floor 1960ish-style machine room, with a row of keypunches, tapes, unit-record gear.

The PDP-1 has a good library of paper tapes of which one is a copy of the original Space War. It has two control boxes, and we fire it up every once in a while and let visitors play.

Anonymous said...

And old VAXes never die...

dhogaza said...

"The first two sit in a room refubished to look like the DEC "Mill" in Maynard (wooden floor)"

Did you soak the floor in lanolin to make it truly authentic?

I never worked for DEC myself, but did consult there (compiler work, used to red-eye from the west coast, take the DEC Air Force from Logan to Parker Street, and from there a DEC van to the Mill).

Friends who worked there pointed out that the lanolin in the floor was hard on rubber soles (it used to be a woolen blanket factory).

A working PDP-1, that's cool. I only go back as far as the "straight 8".

jon said...

You are a bunch of realprogrammers !!!

However, I must admit, Im am another real programmer (perhaps not that real as you... but ... slightly real anyway)

I remember the times I studied my first Fortran course (Fortran 66, of course). Punching cards to get the program compiled by the system managers ... Piles of printed output with compiler and loading errors to get my first homework done...

Some time later, when I told that the old ancient times were that way to my first PhDs student, he simply didn't believe me, until I showed him a real card (a real programmer's real card).

These young people using iPhones can not undestand the feeling when the cards slipped from hands, fell and mixed ;-)

EliRabett said...

One quickly learned what columns 73-80 were for. . . .

John Mashey said...

Piles of printed output.
One class I taught had a blind student. One feature of the IBM 1403 impact printers we used was the provision of a "Brailler" option, that converted text output to Braille, including doing multiple hammer strikes to make raised dots, a feature not so workable on laser printers. Of course, this causes (atl east) a 6X expansion. He was a good student.

But at least once, he got a *large* core dump...

steven said...

Actually for CRU there is a limit.

The limit is 18 hours. I know this because I sent them an FOIA asking for their policies and procedures on FOIA. They responed that they THOUGHT it would take more than 18 hours so they asked me to modify the request.

I found it funny that they found the guidelines, consulted them, and determined the 18 hour limit and then thought I should alter my request because it would take more than 18 hours to find what they had found.

Anyways, I guess I was used to an organization that had a published employee manual. The kind of thing every worker has read and signed when accepting employment. You know assignment of rights, policies and procedures. It's right there on your shelf. I did not appeal my request since I dug a bit more to find the information on the web. Took a few minutes.

Now, of course, the whole fiasco can be avoided by proper record keeping and data management.
Luckily, when I first started to program the people who came before me had nicely organized files.
tapes, punch cards, 9 track, 1/4 streamers. They kept good records because their jobs required it. Maybe because what they were working on was important. But now that I think of it, even when I work on frivolous stuff like various editions of rare texts, I could count on the people who worked before me. One guy who did editions of Dryden ( see UCLA the Dryden project) had racks and racks and racks of 3x5 index cards. each hand written and precisely numbered. Vinton Deering. had editions of the bible on other racks of index cards. It was back up for the computerized version.

And confidentiality agreements? Why they are in the right hand drawer. Under..."agreements, confidentiality. The admin kept a paper back up and an e copy. All in a database. Hundreds of them over the years. The legal department also keeps the original. As time goes by these records are transferred to a document storage facility. Imagine that. Then, 10 years down the road, when I need to pull a record after many moves from building to building, I call up legal. they call the storage guys.
The documents are all in boxes, indexed and numbered. The box is "pulled" and on my doorstep.
But hey, that's not climate science.

I think my friend who does storage also has your dentist as a client and definately your doctor. these people too care about what they do. They are accountable.

So what was Jules' point?

Dano said...

I had presumed I had forgotten the tape. Curses!



Anonymous said...

18 hours per FoI
60 McIntyre proxy FoI requests
13 CRU staff
8 hour day

If I'm correct, that's ~10 days of the CRU doing nothing else but servicing McIntyre's 60 proxy FoI requests!

Now there was nothing wrong with McIntyre making an FoI request, but it is transparently obvious, that obtaining information wasn't his motive, because he could have done that himself with a single FoI.

So why then, did McIntyre do what he did?

It is hard not to conclude from his actions that he intended to achieve something completely different from merely 'obtaining information'. By initiating a torrent of vexatious FoI requests. It is likely he saw this as a legitimate way to pressurise the CRU management - Phil Jones.
It's harassment, plain and simple and it fits in with McIntyre's insinuations and quote-mining of anyone his paymasters don't like.

Random mouse

Anonymous said...

anon n+1

The Democrats', and President Obama's, War on Science Continues

Democrats Harass EPA Scientists with FOIA

" With midterm elections not far off, Democratic opposition researchers are armed with thousands of pages of records obtained from theEnvironmental Protection Agency through the Freedom of Information Act, far outpacing known Republican efforts to pry information loose from the agency, records show. "

EliRabett said...

They caught up