Posts

Posted on

Support YOUR local library

Take the SLAC Library Survey

Not too long ago, the SLAC Library occupied a large space on the second floor of Central Lab [1], the heart of the administrative and research activities here at SLAC. Due to shifting priorities at the lab, and the need to renovate existing space in Central Lab for future experiments, the library was packed up and moved to new quarters. Those quarters are in the lobby of the Computing Center (Building 50). The amount of space lost between the old digs and the new digs is unimaginable and most of the library’s collection is now stored of reach of the average browser or library-goer.

The laboratory assembled a task force [2] to determine the fate of the library. My understanding is that this task force is deciding between several possible scenarios, such as an all-electronic library. My own opinion of this kind of choice is that it is devastating to an entire field of research. Before I go into the next paragraph of opinion, I invite you to take the SLAC Library Survey yourself – note that the very last page of the survey offers a big text box where you can express your own thoughts on why a library like SLAC’s is a useful resource (or not . . . your opinion is entirely your own).

Here is my opinion.

The United States is faced with a grim economic situation and part of restoring the strength of the U.S. economy is to make investments in education and science that train the next generation of innovators so that we can produce the discoveries that create new economic sectors or opportunities. Part of the U.S. scientific investment portfolio is its national laboratory system. SLAC is unique in this system in many respects, primarily because of the openness of the lab combined with its extremely close ties to Stanford University, just two miles down the road.

With the Visiting Scholar program here at the lab, even a lab user can gain access to Stanford Libraries. However, not having a library physically at SLAC does create a barrier between you and your resources. Not only that, but it’s not even clear to me that should all the resources of the SLAC library go to Stanford, there is any room to display them as a single collection for researchers to browse. The SLAC and Stanford Libraries have electronic journal access, but there are many books and conference proceedings available only in print.

I view the resource afforded by the SLAC Library as part of enabling the positive Black Swan [3]. The Black Swan is the unexpected event that completely changes the course of your thinking. It’s the swan that, when discovered, takes the induction that “all swans are white” and turns it on its head. These positive Black Swans are responsible for the general progress of science, and our field is no different. Having the ability to go looking for one thing in a library and find something totally unexpected on a neighboring shelf – something that changes the course of your thinking – in invaluable and currently unreproducable online.

There is also no way to quantify the benefit as too small – libraries cost money, for sure, but the value of even a single discovery enabled by happenstance library browse is priceless. Another way to look at a library is to wonder how many people have been able to determine that an alley is blind, just by looking over past work, before they march down that alley? How much wasted effort, spent reproducing a failed result, is prevented by a resource like a library? Can you quantify that, and put it on an Excel spreadsheet?

Digital publishing has enabled a glut of information to be available at your fingertips. The problem is that, unlike a library, it’s not ordered in a useful way. The library cataloging system has not yet been replaced by keywords, and there is no substitution for the human eye and its ability to browse. SPIRES is nice, but when I use it I find half the time I get too much information and the other half I get too little. Digital searching takes a glut of information and returns a glut of information, while the library smooths out the fluctuations and allows the mind to focus on a new idea, an interesting title, an old undigitized book.

I worry that at a time when innovation is so critical to the U.S., SLAC is considering disposing of an engine of innovation to its research community. When reading “Beamtimes and Lifetimes” by Sharon Traweek, I was struck by her description of the SLAC library. She noted that it was never that busy, that there were people reading or napping, but not too many of them. That level of usage hasn’t changed much since the 1970s, when she wrote her description; yet, SLAC has been involved in major, Nobel-prize-worthy discoveries in that time and, I believe, should continue to do so. I worry that once we cast away a fundamental resource like a library, used frequently or not, we’ll lose even the infrequent browsings that create rare, Nobel-prize-winning work.

[1] http://steve.cooleysekula.net/blog/2008/06/26/i-remember-libraries/

[2] http://today.slac.stanford.edu/a/2009/03-30.htm

[3] Taleb, Nicholas Nassim. “The Black Swan: The Impact of the Highly Improbable”

Posted on

Communication: Breakdown

It’s been a number of years since I had to think – I mean really think – about electronic communication and HEP. On BaBar, I became accustomed to a specific model of communication. E-mail for personal correspondence, for which no record inside the collaboration is necessary (e.g. arranging a meeting time); Hypernews for conversations whose archival value is, at least, non-zero (e.g. the time of the meeting); phone for connecting people together in a meeting (via MeetingPlace); the web for locating presentations either in a central database or a site (PDF preferred, please!).

ATLAS has adopted EVO [1], which was developed jointly by Caltech and the DOE and leverages technology built by the VRVS collaboration. Years ago, a colleague of mine introduced me to VRVS. It barely worked on my linux machine, and I quickly abandoned it. That was sad, because it was built on JAVA and it really should have worked well. I wanted it to work.

VRVS has come a very long way, if EVO is any indication. That said, EVO seems to be the butt-end of many jokes inside ATLAS and certainly it seems to inadvertently induce delays in meetings, due to troubleshooting. The very first day I attended the LBL ATLAS jamboree, I decided to fire up EVO and see what the meeting looked like to people on the other end of the connection. I quickly realized that the tools in EVO allow a participant to exercise a great deal of flexibility in participating in a meeting.

EVO adds extra dimensions to my old collaboration model. Video and audio are totally integrated into the client (Koala), so I can choose to send video or not. Chat is also built into the client, distributed to all the participants in the meeting or directable at only a single participant. Today, for instance, during the Hadronic final states workshop at SLAC there were some problem moments in the meeting when firing off a short chat to the speaker, letting them know there were audio problems, was EXTREMELY valuable. That, and when audio would drop out a few chats fired around to participants in the meeting could help isolate the problem.

I’ve decided that EVO augments my old model in a very positive way, even though it brings inevitable technological challenges. The challenges I’ve experienced so far stem mainly from the marriage of the phone system (literally, a phone handset that someone uses to dial into the meeting) and the EVO internet-based system. The “phone bridge”, as it’s called, seems to introduce the most problems. Feedback, dropped connections, etc. seem to stem largely from the phone bridge.

Based on the last 5 days of ATLAS-related meetings, I’ve come up with a set of personal “best-practices” that I intend to implement when running, or even participating in, ATLAS meetings.

  1. It’s important that the moderator (me, let’s say) focus on maintaining the connection at all times, and sending a “pause” signal out to people via chat when things go badly. It’s therefore also incumbent on the speaker to keep an eye on the chat window for emergency messages.
  2. Having a “shadow” in the audience, someone who is physically at the meeting site but dialed into EVO via video and audio (in fact, even listening to the audio while in the same room as the meeting is a good idea). This person is like the audio engineer, listening to the performance and alerting the moderator of problems. I did this today as an experiment, and I found that if I brought the volume on my headset up to a high enough level (not too high) I could mentally ignore the audio in the room and focus on the audio coming across EVO, even though I was in the same room. It was then immediately clear when the speaker dropped out from the connection, important if they were physically in the room and thus oblivious to the fact that no one on EVO could hear them.
  3. Try to avoid the phone bridge, but if you must use it watch it like a hawk. The Shadow will be of great value in monitoring such a thing.
  4. Wireless internet seems to work fine, but I’m a big fan of hard-lines since the TCP/IP protocols are more standardized and well defined there (there is also no encryption overhead at the network device level). I’ve heard rumors that hard-line connections to EVO are more reliable than wifi connections, and a priori I can see why this might be, but my own experience says it’s fine. That said, I fundamentally trust hard-lines more than wifi.

I plan to keep these in mind while moving forward. I’m particularly keen on the moderator/shadow model. Communication in such a large body as ATLAS is hyper-critical to the mission of the experiment. Minimizing delays is also key to maximizing meeting time, and thus getting out of meetings and getting to work. My collaborators seems to succeed in spite of the frustration they seem to feel about EVO. Me? I like it, even if it does have idiosyncrasies. I like that chat, video, audio, and identity in the meeting are all tied into one service.

With that in mind, I intend to keep EVO up during the day as a means to facilitate cooperation between my colleagues and myself. If you’re on EVO, let me know and I’ll add you to my buddy list on chat. I like that I can keep G-chat and AIM for personal use, and keep EVO for the physics side of my life.

[1] http://evo.caltech.edu/

Posted on

The Tiers of Many Colors

The first two days of this week, I occupied part of my day listening to the presentations and discussion at the ATLAS Western Tier 2 User’s Forum at SLAC. A “Tier 2” computing facility, according to the ATLAS computing model, is one which has the resources available to maintain data on disk for physics analysis, both by individuals and groups. SLAC is a Tier 2, and this was a chance to collect feedback from users about their experience while also telling them about new developments or tools to ease their lives. Things weren’t always in the easy direction, of course.

A few things stood out to me. The first was a very good presentation delivered by a SLAC graduate student relating his own experiences as a user doing physics analysis at SLAC. He discussed topics that spanned the space from getting critical data sets, to using the GRID, to running jobs in SLAC batch. As a newcomer to this experiment, I found this talk at a level particularly engaging for my own needs. Every site is a little different, but I presume that the common set of successes and frustrations likely remains the same at any location.

The other was the sheer challenge to collecting resources at a site so that enough people can do all the things they will need to do to achieve solid physics analysis. This challenge crosses many boundaries – CPUs, disk space, tape storage, purchasing, electrical infrastructure, cooling, and politics. The discussions on this particular issue – resource procurement and management – were especially interesting. The underlying concern in everybody’s mind seems to be that no matter how much you think you need now, it might not be enough come the arrival of real data.