As mentioned here earlier, the last Congress decided to not pass most new FY 2007 spending bills before leaving town, putting these off until the new Congress convenes, and running the government on a continuing resolution, mostly at last year’s spending levels. There is speculation that the new Congress may decide to not even try to put together and pass FY 2007 spending bills (the fiscal year started Oct. 1), instead just funding things by a continuing resolution for the rest of the year, mostly at FY 2006 levels. The Fermilab director Pier Oddone has issued a statement about the implications of having to run Fermilab at the FY 2006 funding level for the rest of the year. These would be dire, including having to take such measures as completely shutting down the lab and furloughing its employees for a month.
This would be extremely bad news for Fermilab, coming at a time when they have been having great success with getting the Tevatron to run at ever greater luminosity. The machine has just set new records for weekly luminosity, monthly luminosity, and initial luminosity (you can follow their progress here). While everyone is concentrating on the LHC, the Tevatron remains the only machine in the world running at the high-energy frontier, and the most likely source of any surprising new information about beyond standard model physics during the next couple of years. It would be a great shame if budget problems were to have a negative impact on this.
I don’t have any information about what the impact of these budget problems might be on particle theory or on mathematics. For mathematics, the impact may not be so great since, after several years of sizable budget increases at the NSF, the FY 2007 budget request for mathematics at the NSF contained only a 3.2 percent increase.
There’s a quite interesting interview in the latest (February) issue of the Notices of the AMS with William Rundell. Rundell was the director of the mathematics part of the NSF until last summer. He describes how during his tenure the NSF emphasized “single-researcher” or “PI” grants, saying that:
If you take any block of time from NSF’s beginnings to now and you ask, what were the best years for the DMS single-investigator grants or for senior researcher increases?, the answer is the period of 2001 through 2005.
Rundell notes that during this time the number of grants went up by only a small amount, maybe 10 percent, but that the value of each grant “went up enormously”. Before 2001, people were being given at most one month of “summer support”, now junior people get two months, and senior people often a month and a half or two months. While inflation and average university raises have been around 2 to 3 percent, the academic star system has had stars (the people most likely to be getting these grants) receiving 6 to 7 percent raises, 10 percent promotion raises, and big hikes in salary when they move. So, the bottom line is that a lot more money has been going to a small segment of the mathematics research community.
The interviewer states that “Most mathematicians believe PI grants are the most important part of the DMS”, but I wonder whether that is really true. Rundell also explains that the current system leaves most mathematicians with not much motivation to lobby for an increased NSF budget, especially if most of the increase is going to go to a small number of well-paid people:
I think it is probably true that the mathematicians who get the money aren’t pulling their weight for justifying us to get more. And on the other hand, those people who are disenfranchised have no incentive to do that.
Personally I’ve never understood the logic of devoting such a large part of the NSF research budget in math or theoretical physics to increasing the salary of the best paid people in the field, although I hear that once one achieves such a status the reasons become much clearer. Besides the “summer salary” though, these grants do fund many things that are important for the health of university math departments, especially supporting graduate students. Rundell claims that over this same period the NSF has doubled its support for graduate students. This is probably reflected in the data contained in another article in the new Notices, an annual survey of new doctoral recipients. This survey finds the number of Ph.D.s awarded last year in mathematics to be 1245, the highest number ever recorded. Four years ago this number was at a local minimum, with 948 mathematics Ph.Ds awarded.
Also supposedly suffering from funding problems is the high energy theoretical physics group at Harvard, where, according to one of its faculty members, because of feminism the university has been unable to afford competent computer support. As a result the group has recently had to shut down its web server (schwinger.physics.harvard.edu), and evidently has had several of its machines broken into, with no administrator around to deal with this. There’s a huge on-going problem with university computer systems which seems to be the same thing that happened at Harvard. Many groups of hackers have broken into a large number of insufficiently well-protected university unix systems, often installing trojanned versions of the SSH software. The trojanned SSH client programs then gather people’s usernames and passwords as they are typed in when SSH is used to login to another system. These are used to break in to yet other systems. Since SSH is the fundamental tool used to manage logins between different machines at most universities, this is a very difficult problem to deal with.
One reason I’ve mentioned this is to warn people to be very careful about using SSH, especially using it to login from a system not at your home university, since the SSH program on the machine you are using may be trojaned. Better to use your own laptop, with its own SSH software. I’d like to discourage posting of comments about computer security here, since most such comments just spread misinformation of one kind of another, just making problems worse. There are many other places on the internet to get information about and discuss these issues.
Update: There’s more about the 2007 NSF budget here.
Update: Today’s NY Times has an article here. It seems that many other labs, including RHIC and Jefferson Lab, are facing similar problems.
just regarding ssh. Try using private/public key (DSA for instance) authentication when possible and most of those problems go away. And don’t login *from* a computer you don’t trust (even when using keys)
If LHC finds nothing people are expecting (SUSY, higgs) nor anything completely unexpected (my favourite) then budgets will get far more stringent in the future.
“Congress may decide to not even try and put together and pass FY 2007 spending bills”
“try and”? Shame
OK, OK, Grammar Nazi,
When I proofread the posting I did think that the writing was worse than usual, but also that I was too busy to do much about it. And, hey, standards on internet blogs are low in general, right?
Simply put, I do not believe Fermilab’s Pier Oddone. Inflation is about 3%/yr. Thus, if 2007 funding is at 2006 levels, they will have to cutback by 3%, not close down for a month (an 8.3% decrease).
Besides, when was the last time Fermilab produced any good work? 1976? They have hardly established a good track record in the last few decades, and better they live within a budget than that my taxes go up.
Kevin,
Oddone gives a detailed justification in his letter to the DOE. This far into the fiscal year there are strong constraints on what he can do to save money, especially since supposedly there are no major capital spending projects going on that could be stopped or slowed down. His estimate for what they can save by shutting down the lab for a month is 3 percent of the budget. You’re not going to get an 8.3 percent savings since what you save by turning off the lights and furloughing people is only part of your operating costs. You’ll continue to need some people to stay there, and you morally can’t do things like shut off medical benefits to your furloughed employees.
If you really don’t think Fermilab produces any good work, you shouldn’t be arguing for saving $11 million by shutting the place down for a month, but for saving billions by shutting it down permanently (along with the rest of US HEP research, which hasn’t done any better than Fermilab). And please, everyone, take the aggrieved arguments pro or con about taxation levels to other blogs hosted by people willing to tolerate them.
In defense of Farm ‘n Lab…. I agree that nothing earthshaking is
has come out in a long time. An important consideration, however,
is that there are standard nuts and bolts measurements, such as
more accurate values for masses and widths, soft-scattering
measurements, etc. This is physics, even if not the exciting kind.
I have a question for experimentalists or phenomenologists
who might be reading. Will measurements at the Tevatron be
useful for interpreting LHC data? I think they should, since
the background at the LHC will be hard to subtract otherwise.
Someone else would certainly know for certain.
If Grammar Nazi is reading, there is no need to chastise me. Re-reading
my last sentence fills me with shame.
Kevin:
1976? For most of us, the discovery of the top quark was certainly good work. Perhaps you meant 1996? In any case, the Tevatron has put lots of limits on new physics, and has precisely measured a lot of Standard Model parameters. It also does good B physics.
Grammar Nazi:
“Try and” is perfectly grammatical in typical English usage. Surely you have better things to do than carry on a pedantic crusade based on arbitrary rules? For instance, you can go to this post at Language Log to see “try and” being used by a coauthor of the Cambridge Grammar of the English Language.
backgrounds have already been measured, and will be much worse at LHC.
Observing tops, Z, W… at LHC will help in calibrating LHC detectors.
People moving from Tevatron to LHC will bring their expertise.
The main goal of Tevatron was and is discovering something new just before LHC. This would have some scientific value, because Tevatron detectors and backgrounds are already understood. Nevertheless it seems to me an example of poor competition: Tevatron runII is not really needed for physics and needs a lot of resources. For comparison, CERN preferred to close LEP (despite the hints for a higgs, leaving the next strike to Tevatron) and go on with the next big project. Hopefully this will turn out to be a good and brave decision.
The more I read your blog the progressively more skeptical I become about a meaningful career in hep; maybe ‘financial engineering’ isn’t such a bad move to make if you want a remote chance to make a decent living *AND* use all the new tricks and skills you pick up during your studies.
How say you, Pete?
I am not any of the Petes, but I have done a fair amount of Financial Engineering after doing a PhD in HEP theory. What you need to be a Financial Engineer is knowledge of probability theory and the ability to solve PDEs. Also, programming skills are essential. However, you are unlikely to be using any techniques you learned after your first degree. The reason that employers choose PhDs is that there is hot competition for the jobs and PhDs look better on paper.
Peter – do you think there is a niche for physicists/mathematicians with professional-level computer knowledge to run these university systems? I would think such a person could better serve the needs of physics/mathematics faculty.
-drl
drl,
I do think it is a good idea for physics/math departments to consider having their systems managed by a permanent person with a Ph.D. who teaches and has some sort of faculty position. This has worked well here in the Columbia math department, with me and my predecessor (but then I would think that..). The main problems I’ve seen with departmental computer systems occur when one tries to run them with grad students or poorly paid/part-time computer administrators, who don’t stay in the job very long. If you want a reliable computer system, you need to have someone smart in charge of it who is around long-term, and has the authority to enforce standards and spend the money needed to ensure reliability. The security and spam difficulties associated with running mail servers and multiuser unix systems are huge, and the day is long gone when a faculty member or grad student could set these things up and get them working reliably in their spare time.
Michael,
I think Chris is right that most “financial engineering” jobs don’t use that much sophisticated mathematics. But they certainly are much better paid, with much better job prospects than in academia, especially if you want to live in a place like New York or London. At the moment half of all personal income in Manhattan comes from the financial industry, and extrapolating from past trends, that should soon approach 100 percent. Not clear to me what the future holds…
Mathematics is a perfectly healthy field, theoretical physics much less so, with the problems I discuss ad nauseam here. Experimental HEP in the US has serious questions about its future. For both theory and experiment, the future should be a lot clearer 4-5 years from now after initial results from the LHC. If the LHC doesn’t find something that disagrees with the standard model, experimental US HEP will be in big trouble, theoretical physics will also be in huge trouble if it continues down the same road it has been following.
Hi Peter Orland,
“Will measurements at the Tevatron be useful for interpreting LHC data?”
“Tevatron-for-LHC Report: Preparations for Discoveries,” hep-ph/0608322, has some information about this.
Regards,
C
Peter, sorry for asking a question about computers, but maybe you are in a unique position also for this issue. I would like to understand if forbidding direct ssh access (to everything, including the http server) is an exaggeration. To users like me it seems more annoying than having some hackers, but our professional administrators have a different opinion, and, being not competent enough, I cannot understand if they are right, or if they like to have a too easy life.
C,
Thanks!
ssh?,
Unfortunately, at the present time there is a good case to be made for severely restricting people from logging in to accounts that can run a shell from machines that one doesn’t trust. There are just too many compromised machines out there, and if someone logs in from one of them, a group of hackers now can and will get into the account. From a shell account, even if they don’t know much they can cause significant trouble. If they are very knowledgable they can probably get root access and cause huge amounts of trouble (compromising many accounts on the system, trojaning ssh so they can get into other systems, trojaning the system software so that the server needs to be completely rebuilt, etc. etc.).
Security is always a balancing act between making it convenient or at least possible for people to do things over the network that they need to do, and protecting systems by minimizing the number of potentially dangerous network connections. There’s no easy answer to this…
“I do think it is a good idea for physics/math departments to consider having their systems managed by a permanent person with a Ph.D. who teaches and has some sort of faculty position.”
That wouldn’t work. Having someone who teach as system administrator “when they have time” only means the systems will basicly be left unmanaged and the only thing they end up doing is add or remove a few accounts now and then.
It’s a full time job trying to keep up to date with security updates and especially trying to anticipate any possible future attacks (then again, just avoid windows and you’ll automatically avoid most of the security problems of today).
Ari,
“That wouldn’t work.”
Well, my experience is definitely that it can work. My colleagues here in the Columbia math department seem to be quite happy with how the computer system runs, and it’s basically run by me, at the same time that I teach one course per semester (some semesters a seminar for the graduate students giving them practice teaching, which requires little preparation). Before I took over this job, it was done on a similar basis by someone else, and people were also happy then with how it worked. It would not work if the person involved had a much larger teaching load. It also can take quite a few years to get a computer system stabilized, secured, and working in a reliable, secure and easy to maintain manner.
And no, avoiding Windows is not the solution to all problems. The problems the Harvard theory group seems to have had which required shutting down their systems had nothing to do with Windows. And Windows/Mac/Linux arguments and evangelism are completely off topic. Don’t even think of it…
Just to add to my previous post, I use them all. There’s different sets of problems each one solves the best.
Regarding computer and network security see this sobering article:
Attack of the Zombie Computers [botnets] Is Growing Threat
(NY Times, 1/7/2007)
(Windows XP Home Edition, not to mention Windows 98, has for years been considered a botnet disaster waiting to happen.)
Ok, Peter’s idea to take discussion about security issues elsewhere sure is a good thing:
1) “trojanning” is a wrongly spelled verbing of a bastardization of “installing a trojan horse program”. What happens is that the SSH client is either replaced or “debugged” by something that captures password information. Thus, “compromising the SSH client” is the phrase to use.
2) Plank says: “Try using private/public key (DSA for instance) authentication when possible and most of those problems go away.” They do not and DSA is NOT a recommended algorithm for public key cryptosystems. However, “two-factor authentication” will alleviate some problems.
3) Ari says: ‘Having someone who teach as system administrator “when they have time” only means the systems will basicly be left unmanaged’ Ab-so-lutely. It’s fun to do half-assed system administration from time to time (don’t we all love to tinker) but how can you ever dot the i’s and cross the t’s on a network with more than a handful of servers, workstations and users ? This is a high-dimensional space and the inevitable random walk will quickly take you into the regions that are labelled ‘here be downtime’. Not to be insulting but I have noted that often people are somewhat optimistic regarding the true state their network is in, especially if they have an academic background. Ah the good old times when all our data was on a single harddisk under the teaching assistant’s desk (no, it was not in the backup pool either).
Also…where’s the documentation?
4) Two words: Tripwire or Radmind.
Bu enough of this, enjoy the Sunday.
“trojanning” is a wrongly spelled verbing of a bastardization
It’s interesting how wrongly spelled verbings of bastardizations can be absolutely clear and meaningful. “Trojanned ssh” simply says it all in just two words…
I can speak to life sciences. An NSF proposal on which I was PI scored in the top 15% was declined; less than 10% of submissions can be funded at present (in recent years around 20% have been funded). Fortunately we can scrape by for awhile on other funding sources but for others this will be the difference between surviving and not.
Kevin typed: “Simply put, I do not believe Fermilab’s Pier Oddone. Inflation is about 3%/yr. Wellllll… belief != data, and the Consumer Price index != inflation of scientific equipment and supplies (which are increasing faster), or of the cost of electricity required to run the experiments. Moreover, the cost/time of running a scientific experiment is not constant, even in constant dollars. In addition, nothing in the above post indicates that Fermilab itself is going to see a 3% cut. It is common, even when agency budgets are flat to implement across-the-board cuts that are considerably more severe.
In other words, it is obvious that you are operating in a severely data-deficient state, and that your beliefs, while strongly felt, are not based on much beyond strong feelings. Finally, I would note that the Republican congress and the President were planning to DOUBLE the NSF physical science budget over the next several years, and I doubt that the Democrats will scuttle that particular plan. So you might want to avoid getting too excited about a one-year 3% cut. Over the longer term, what you want is simply not going to happen.
OK, OK, will fix the bad spelling…
I suggest
trojan’d
the double n doesn’t look right—it puts the stress on the second syllable
but neither does trojaned.
Verbing is just fine in this case.
So how much funding for the Tevatron would be required to keep it running until the new budget is cranked out ? I’m guessing on the order of ~500 M$$…..
Perhaps Jim Simon could talk to Bill Gates, Steve Ballmer, Matt Lesko, Sir Richard Branson, and other high-tech philanthropists, and they could in turn network to their peers….shortly, in a geometric progression, the bread might be made available if these parties were made aware of how dire the straits are….?