I have spent an inordinate amount of time getting my computer to be useful at my new position. This is a surprising revelation, given that a central part of my role is to develop new software and employ new approaches to science, collaboration, data analysis, and automation. These are all things I'm really excited about. I sort of assumed that "open research environment" meant that researchers could, within reason, configure their computers how they see fit.
This is, of course, simply too much to expect from a federal organization.
IT, and the freaking huge agency that my huge agency is a part of considers computers as boxes that deliver email, intranet forms (via lotus), and run various microsoft products to bureaucrats so they can move money between accounts and sign forms. As a result, a massive list of changes to the default windows installation is applied to every box, be it windows, mac, or linux, under the guidelines described by the FDCC (Federal Desktop Core Configuration). This list generally makes the computers more secure, moves account control to active directory, and enforces a strong deliniation between user and administrative actions on each computer. (something that windows 2000 and XP did not do well out of the box). The upshot of this is that computers do not go down very much due to people downloading virii with their Lady Gaga mp3s, because they can't install file sharing apps, and they are relatively immune to easy infection vectors.
The downside is that researchers, programmers, and engineers are effectively nerfed from their computers. Given the number of actual research scientists in a nominally scientifically oriented organization, you would think there would be a coherent policy on this, but that is not the case. There are numerous exceptions that can be granted via a maze of signitures and paperwork.
I do not think IT has any idea of what engineers and scientists do. I have nagging fear that people on the outside have vague thoughts of test tubes and slide rules.
Thus, I have to beg and plead for administrative access to my machines. This makes it very hard to try new things to see if they help solve problems better than the old things. Explaining to a helpdesk person what Mendeley, Zotero, or Simion is and "why it is needed" is a major waste of time.
My heart hit the floor when i called the helpdesk to ask them to install Labview. "What is labview?" the tech asked, incredulously. Labview must sound like a webcam program. I have had a lot of explaining to do just to be able to start to work.
The paperwork has been filed for administrative access to my desktop. Hopefully, this will lubricate the rails a bit.
It's a brave new world, kids.
I stole an IP address today. It was one that was previously being managed by the DHCP server. I'm not altogether convinced that it was necessary. I had set up a linux server this week and then loaded the intended application on it. The application at one time or another does a reverse ARP to back translate its IP address to a hostname. The name that comes back is whatever the DNS server thinks the port number's name is that it knows the port by and not the hostname. Every now and then it changes the IP address under DHCP and the name changes. The application expects to have config files that are named to correspond to the host name as is returned by the Reverse ARP. So having some obscure name returned is not helpful.
ReplyDeleteI had been seeing this symptom for months in X-terming to a remote server. The login scripts attempt to ARP to get the IP address of my local computer so it could set the current DISPLAY environment variable. The hostname it has is the internal port name which does not back traslate to an IP address. So I always have to manually set the DISPLAY value.
Its all nuts.
On the other hand, while in the WINDOWS world we are severly restricted by corporate IT, in the linux space we seem to be operating as though completely transparent. Linux? What's that?
Xming, I'm thinking, Xming!
ReplyDelete