Wednesday, February 22, 2006

Yikes! A horrible mistake!

One of the researchers that was participating in the paleoclimate meeting contacted me and pointed out a horrible mistake that I made. I did not give proper attribution for the posts for each of the authors. I merely included the whole agenda in my first live blogging post. He pointed out that this is his (and cocollaborator(s)) IP and that not attributing was effectively theft.


*THAT* was *NOT* what I wanted to do! I have asked for forgiveness and I am planning on either removing the offending posts or adding atrributions as per his preference.

Any of you that get pissy about that, can fscking BITE ME. This is HIS research. I didn't atrribute as well as I should have and he has spent a helluvalota work on this.

I'm awaiting his judgement as I speak.

Sick and wrong irony was this was the researcher I really wanted to talk to about the SC06 and SC07 challenge shots...

Talk about a bad first impression!



Anonymous said...

While I'm a little unclear why informal running commentary on a scientific meeting requires proper attribution -- if I put a dollar in a jar every time I've informally said "that thing that guy wrote" or equivalent instead of the proper cite, I could buy a new car with the proceeds -- I have no problems with any emendations or removals you decide to make.

Is there anything on my link page that you would want changed?

Will Baird said...

No, you're fine, Carlos.

After a lot of thought of what happened n3ur0last night, I am more pissed off than anything. I happen to agree with you: completely. This was a commentary and a handful of notes. The agenda was the very first freakin post and since this was merely a set of notes and I am NOT a climate researcher, this guy went off the handle for no reason. Contacting my fscking BOSS?! Not to mention with an accusation of IP theft...that's poison where I work.

A polite note would have been fine and I'd have been delighted to make changes.

My boss was more amused by the end of the day than anything. My job is not research. I am an enabler and that allows me to help along lots of researchers and that allows me to sample and participate in a lot more...if I am proactive. That was my attempt to do here. The 'tard involved here was the one I wanted specifically to try to work with and essentially get (if all worked out well) lots of free CPU cycles: as in millions, potentially, of MPP CPU hours.

The guy I worked with this year is writing up a paper for Nature on modeling hurricanes based on the cycles he got from me. He originally made his career off of free CPU cycles for climate models and this paper is going to boost it more. This new paleoclimate guy just lost a nice chance here. This new guy emphatetically not my intended collaborator anymore.

It means that one of the eras that I find most fascinating won't be modeled for this new two year project I want to do, but...I'd rather work with someone that wasn't quite so...well...y'know. :D

I may scrap the idea of working with the paleoclimate guys at all, but we'll see. Give me a week and let me cool down.

Anonymous said...

Well, you know that tag of Yeats (on Ireland, originally): great hatred, little room. I can't imagine that modeling paleoclimate is a very large field.

His loss. The operative word in "charm offensive" is not the second one. Although maybe he had a really bad day or been burned before or something.

Will Baird said...

The funny part is that I think I have the lead on the new system we're buying that will have a sustained system performance (SSP)[1] of over 10 teraflops to work on the project that'd be giving out CPU hours.


Dedicated time on one of the fastest machines in the world, flushed down the toilet. I really hope the guy was just having a bad day. I really do. His stuff is fascinating.

Oh well, in the words of the hooters girls that taunted a friend of mine when he declined a date with one due to his recent engagement: too bad, so sad, look and see whatcha could have had. ;)

1. Most systems like to chat up their peak performance which is often over 150% of its benchmark performance and often (with current architectures) over 20 better than real world apps. SSP is our 'new and improved' descriptor. So, if the SSP vs peak holds true, this new system will have the peak performance around 200 teraflops. Using the full system as we would for this project would give millions of CPU hours away...