Wednesday, December 1, 2021

Spoke and hub

In one of the Cal Newport books (I’m re-reading his last three to cull all the insight and distill, more on that probably later), he mentions the setup at Bell Labs and a famous MIT building. They all have a “spoke and hub” setup so people can collaborate and interact serendipitously and then also withdraw in places of privacy and focus. 

This is a key metric I feel for a good scientific environment. And maybe if you are looking at a place of future employment: when being whisked around the campus or department, can you identify where the hub is? Do people seem to appreciate and respect their spokes? Or is it a place where you get your real work done at the coffee shop? 


This sort of hit home with me how we work and how the spaces I’ve occupied as a scientist varied in their approach. Space Telescope, the Kapteyn Institute and Leiden Observatory all had a coffee spot with both free (bad) coffee and space to mingle, chat, and collaborate. ESA was interesting since it was a small department in a huge organization but there was a similar break room. UCT had a cozy break room as well. 


The flip side is how well the offices are appointed and how much people’s space is respected. Here the story is varied. UCT and Leiden were not great as various people felt it was ok to start a sentence in the hallway and expect an answer from me by the time they actually got to my desk and I was still taking my headphones off. ESA was amazing as quiet working on stuff was expected and encouraged. STSCI was somewhere in the middle with open doors everywhere but generally good boundaries, respected by everyone. 


So for those of you on the job market interviewing. This is something to look for. Can you pick out the hub? How do people treat the spokes? Indeed this says a lot about how much good science you’ll get done there. 

Tuesday, October 12, 2021

Technological Determinism and the near demise of IRAF

I’m reading the book “a world without email” by Cal Newport (it’s probably an aspirational title) and he introduces the term “technological determinism” ie when a piece of new technology steers human behavior rather than the other way round. Often in unpredictable ways. Unintended consequences of new technology adaptation. 


Of course there is a corollary in astronomy: the discontinuation of python 2. 


Python 2 is what kept the IRAF tools pyraf alive for many of us. The pyraf wrapper allowed us to simply call a bunch of trusted tools. No need to learn how to manipulate images as numpy arrays, imcalc will do this for you. But then the python project decided to discontinue support for python 2 and everyone had to move to python 3. Grudgingly I did so too. 


This happened while the astropy package had not yet quite reached the maturity of toolset as IRAF. In some ways it still feels like it is catching up (I'd like to rotate this image around this RA and Dec, how do I reduce this spectrum etc). 


OTOH python 3 and astropy were much more optimized for database manipulation. So student projects and the science as a whole much more became focused on that. Part of a drive away from individual sources and more the study of populations from large surveys, accessible from servers. Etc etc. 


Not a bad thing. And mostly something that was happening anyway. But I am left wondering what else it drove. 

Tuesday, June 8, 2021

The zero sum game

Just finished the book “the sum of us” by Heather McGhee. It is a fascinating book about the effects of racism and the adverse effects on white people through the razing of community resources (she starts with the example of public pools and how the response to desegregation was to drain the pool rather than share it). Basically racism is why we can’t have nice things. 


The main premise in her analysis of the thinking of white society in the US is the cornerstone of “zero sum”. If “they” (you’d be amazed how people can put additional baggage in this single word) get more then “we” (same) must be getting less. 

This suddenly snapped a whole bunch of things into context for me. 


This got me to thinking how the “zero sum” attitude pervades academia and astronomy. In some cases there looks to be indeed a zero sum (telescope time, tenure track jobs, and grants). But is there? The zero sum attitude is something that rears its head more in times of (artificial) scarcity: only a few high-z galaxies, just a few exoplanets. Yet it’s pretty clear that the sum of all of us produces more and better science. Inclusive collaborations that produce public data for all to enjoy and use. 


But if you are raised in the zero sum thinking (white, male) then such cooperative efforts are at best a learned experience and at worst threatening. 

Personal and Professional knowledge Management System(s)

I was listening to the “focused” podcast and then an episode of “Mac power users” about PKM (personal knowledge management) systems. Basically what app to use. 


A while back I got into Evernote. It was elegantly designed and allowed met to access my notes on stuff everywhere (iPad, phone, computer). Ideal. I have taken copious notes on a conference for example. I’d like to take notes on talks and tag them (this is aspirational, I still take notes on paper). 


And I like to take notes (plenty of sketches) while I talk to my students. The iPad and Apple pen are really nice to use there. But Evernote isn’t really great with the sketches. And I’ve noticed I stopped using it almost completely. It loads super slow on phone and iPad so I move on. 


So I noticed I partially stopped using Evernote but I do keep notes and save links. But they are scattered everywhere. Two email systems (work and personal) hold notes I’ve taken, there is the iPhone notes app that I’m typing in right now, Twitter and Facebook bookmarks, random text files on my desktop, paper notes, google docs, and I suspect there are more. 


So I need a single system that

1 works on iPad, phone and computers 

2 allows me to sketch

3 synchs easily between everything

4 doesn’t take forever to load

5 OCRs my sketches for ease of finding stuff

6 allows me to share a note easily

7 allows me to back and forward link notes to see connections 

8 can link to other stuff (like references in bibliography) 

9 I can just drag a pdf and other files in and done. 


So do I reinvigorate my Evernote? Or move everything into Craft? Or will craft slow to an unmanageable crawl as I load all the stuff into it? Time to explore. 


This weekend was the final straw. I tried to put some loose ideas I had while running into Evernote. The phone app just sucks. It takes a literal minute to load (yep. a whole 60 seconds, waaaay too long for this). And copies things weirdly, does not capture tweets (critical since so many people tweet links to interesting papers or data etc).


I am migrating notes from *everywhere* to Craft. It does not OCR yet but let's see. 


Friday, May 28, 2021

Data as leverage

The latest astronomy arxiv thing (its a thing. More than a little for many people clearly. Check Twitter.) got me thinking about the use of data and especially reduced data as leverage. 

In an ideal world we all out our final data products on a repository. Documented and free for all to use. Reduced data represents labor. Often a lot. Archiving it properly and documenting it sufficiently is even more labor. So an underlying idea is that people who performed that labor get rewarded for that somehow. A rather central tenet of observational astronomy I would say. 


But I notice some more about this particular case. The labor was probably students and Postdocs. The telescope is private and only *very* recently made any effort to be closer to public and accessible (basically a newly onboarded partner insisted on it). And I suspect this sort of using data as leverage (and thinking in terms of leverage) goes hand in hand with private telescopes. 


I can point to publicly funded telescopes as a safeguard to this. Similar senior people that had a large survey in hand could eventually not use it as leverage. The data was public and a completely different group ran it through the reduction pipeline. 


Now this is not the main thing coming out of this. There is the whole dynamic on senior people not protecting their junior colleagues again. There is the relived trauma of seeing a certain name flash by on arXiv. But it was all made possible by the fact that data is leverage. Huge part of the power play in US astronomy. As if I needed another reason to be an ``open skies" astronomer. 


Monday, February 15, 2021

Post-Pandemic Academia

 “When this pandemic is over....”


I have started to think about what will be different in the post-pandemic university. Some things will be good, some bad. Here are my predictions. 

Paperless is here.

Everything had to be done over email so all forms etc had to be signed electronically. God I hope we will never go back to paper forms that are signed with a pen. I will refuse to do that. Just. Nope. 
I could not be happier. So much easier to file and search. No more crappy yet bigger file size scans of a “signed” document. Adobe had this feature before “The Matrix” came out. Sign your documents like a grownup. 

Recording lectures 

Making recordings of your lectures and PDFs of the slides is going to be standard practice. I’m ambivalent about it. It may be good for student participation and success but also is an indirect additional burden on the professor. Not a huge effort per class but it adds up. But universities have invested frantically during the pandemic in streaming and recording equipment and they’ll want a return on this. 

Hybrid outstays it’s welcome.

Similarly, there will be a push to keep doing classes in “hybrid”. That’s a much bigger effort on the professor’s part. I can’t really do it. Can’t teach effectively to people in the class and online. Just...no. 
But I suspect this will see a push because you can charge more for it or advertise it to potential students. And you can keep doing it in snow days! Oh your kids got a snow day from school. Too bad because you don’t! 

WFH is the new standard. 

Oh I love working from home on writing. Honestly it’s even better in a coffee shop. Or an airport. My office on campus ranks someplace below a busy intersection with McDonalds wi-fi in how conductive it is to writing and other deep work (coding). Erratic temperature control, that weird smell, that time they flooded it, constant interruptions, loud people outside the door, planes landing nearly on top of it, brownouts. Nope home office it is! I already blocked off a day for errands and writing in the before times and in all honesty, I expect I’ll up that to 2-3 days a week. The trick will be to group those things where I have to be on campus for. 

Meeting with research students over zoom

Honestly I’m keeping this. No one has to come to campus and really it’s not that much better in person (see campus office above). I’m thinking individual zoom and maybe group meetings every two weeks or so on campus. 

Poor places replace colloquia with zoom zolloquia 

Yes! Speaker from Australia? No problem! Speaker closer by? Maybe in person. Record these and put on a YouTube channel. Online talks are here to stay. 

No flying for a committee

No TAC or funding committee is *ever* meeting in person ever again. Just nah. So much easier on zoom. When we’re all on zoom. A reasonably well run meeting does not need to be in person. Shame. Too bad. But I also thinks it helps with being more impartial and fair. This makes it easier for anyone to be in "the zoom where it happened" i.e. decisions are made. 

Online conferences are a thing now

And they’ll stay that way. Learn more about stuff while never traveling. Bliss. Seriously. Conference travel can be cut in half. When done well. The social aspect we all miss but they are great for mix and match, asynchronous viewing. Which brings me to...

Conference proceedings are dead

Long live the archived talks and slides. No more overpriced books years later. Good riddance. 

Friday, February 12, 2021

Professor Starter Kit

I am coming to the end of my tenure track period at UofL and part of that is making me evaluate my systems for professoring. I’m also talking to students how to set up your research. So inspired by all that, here are the things to set up before starting to professor. 


Personal Knowlege Management System

This is a first critical setup. Pick your system wisely. I had Evernote already in use partly because it has mobile app as well. Something that allows you to “capture” ideas for student projects, notes on telecoms or meetings and is searcheable! One spot for it all. One system. I keep hearing about others. Omnifocus or Roam Research. It’s worth spending money on. I use Evernote for all this but the phone version is sloooow. And some notes I could have sworn were there are not. As soon as you don’t trust a system anymore, chaos descends. 

A university is a large organization that hands you pieces of paper to ask you about much later. This is the system where scans of all this dead tree nonsense is going to end up in. Grab, scan with phone, stick in PKM and shred. 

Tenure Box

BEST piece of advice I ever got. Make a Tenure box folder in email. Grab a box and write “tenure box” on the side with sharpie. Anything that could potentially make you look good, stick it either in the email box or the cardboard box. No need to sort, you can do that later! 

Email

Inbox 0. This is the way. It’s the only way. Every possible request and task will show up as an email. So it’s best to distinguish between a few timeframes where you file ongoing email assignments and “done” (archived). Mine’s are today/this week/this month/someday/waiting. 

The waiting one is for stuff you’ve done your bit for but are waiting on others. After a while, go through that and send reminders. 

It does not need to get more involved than that but I file email more. Especially useful are folders for each grant application and each grant administered. Grants management at a university is chaotic at best. You can be the one providing structure. Optional folders are for each research student you have and each class you teach. Or not. 

Email is asynchronous. Unfortunately most of the University’s management will think of it as instant message or slack like, expecting an immediate answer. Students sometimes too. Disavow them from this notion immediately. Work email is from 9-5 on workdays. You can expect an answer within 24hrs if it reaches me during those times. IM it ain’t. 

State what action you need. I wish more people do this. The US army has a method for that: BLUF, bottom line up front. In brackets State what you need from the other person. So many academics come up to it sideways in a 3-page essay. The bluntness and utilitarian nature of bluf is appealing to my Dutch self. More of that. Start doing it yourself and ask others to do the same. The number of times I had to parse through several layers of forwarded emails with FYI to realize there was something I needed to do... rude really. 

Email is suck a timesuck that I started calling it “the churn” after the expanse. It still is and can be anxiety inducing but an organized inbox lowers anxiety levels substantially. Also: “meh I’m too late for that.” is also “done”. 

Task Manager

Email is NOT a task manager. Or tracker. You list for yourself what you need/want to get done today. I did paper todo lists forever and they are both demotivating and easily lost. I am trying a Bullitt journal and I’m liking it a lot. I’m working from home though so it’s easy to have this as part of the setup. There are electronic ones for todo lists. Fine. Pick one and stick with it if it works. 

Research structure

Parse your research ideas into projects for (a) you (in your rare spare time), (b) Postdocs and grad student projects and (c) undergraduate projects. 

The latter desperately need structure. So chuck every idle though with the latest data release from a survey into you PKM system. You’re gonna need those. Anything that just requires someone to read less than half a dozen papers and comes with a CSV file of data. Set up a dummy undergrad project: overleaf doc for the report, fileshare for the data, code, and papers. And a setup of the goals. 

Undergrad projects are a net time loss in most cases. The goal is not a paper. But with enough organization and clear goals you may come close to breaking even (it would have taken you just as long on your own). And let’s not compare here with professor-at-full-speed. 

Teaching and Service structure

These are time sucks. Teaching is fun and a primary responsibility and inspiring students is so rewarding yadda yadda. It’s a perfect gas of efffort and time, taking up every minute you’ll let it. 

Odds are that your colleagues don’t care. And underestimate the amount of time it takes to set up a new course. And all your courses are new. So ask about which of your friends has taught the course. Grab their stuff. Steal their slides, their assignments etc etc. 
Once you get there, see if there is a teaching track prof who taught this class. Get their take on the class. Buy them lunch. Steal their stuff. 

Service is often “light” because you’re on tenure track. But it’s considered “light” by those who have been at the place in question since forever. No one left a record on how to do the thing you’re supposed to do for the department. So you have your Task and since you’re the new guy it probably officially Sucks. Parse out what needs to happen. Pick the brain of the last victim and then write it all down in the order it’s supposed to happen in a google document. 

Your future self or the next hapless Service victim will thank you. Share. Make their life easier. Easiest favor ever. 

Grants Management 

Abandon all hope, ye who enter here. Another part of the University that will get mad at you for “not doing things the way we always do them” despite never telling you how to do them in the first place. Tip 1. Identify the grants you really want and identify a (tiiiiny) grant you don’t give a shit about. Do the tiny and/or bullshit grant proposal first. This is your tryout of the grants management system. How long does everything take? What’s a good padding before the deadline? Which parts of Grants Management are responsive to email and which only to threats of physical harm? How many different versions of the proposal clearance form does it take? Who gives you this year’s budgeting numbers? Build a note with how to do what and when. Use this forevermore but be prepared that GM will get revamped and renewed into something even more Byzantine and you’ll have to adapt once more. 

If you’re feeling animosity towards the GM people and their litanies of why laptops aren’t a real computer or why a student can’t have one, remember they are already working with the University and Funding agency’s budgeting management systems every day. They already are in hell. Possibly for years. Take the time to let them tell a story. Patience, politeness and a kind word will get you far far more than impatience and frustration, even if that is what you're feeling in great spades. 

You’ll identify the one person that actually knows what they are doing. This is your new friend. Ply with alcohol and gifts. Bribe outright if you have to. 
GM isn’t there to help you write a grant. This was a major realization of mine. They are there to ensure that the University gets its cut from your grant. Nothing else. 

Routine

Academia is chaotic. On purpose or not, it’s where all the noise of a large organization comes together with all the organizational skills of a bunch of people that never took even so much as a class on spreadsheets. So you need structure. 
Are you a morning or evening person? Do you have a family or no? There are a LOT of variables here and basically none are your employer’s business. You need to decide. With a family it’s pretty much proscribed: you need to get them to school. That sets the starting point to your day. And it’s finish. Do not let others pressure you into working long hours that you’re not optimized for or do that stupid burning of the midnight oil. Business hours are 9-5 people. Workdays from Monday to Friday. I took public transport for a while and it provides a wonderful excuse: “I need to catch the bus sorry.”
Decide when your day has a beginning and an end. Decide when they are. Or it will be decided for you. And then a STEM professor will end up telling you to work 200 hours a week because they’re so fabulous at maths. 

Ranting aside, by having a routine (check astroph, write for an hour) you can make steady and relentless progress towards the stuff you’ll be judged on. 

Design your rest and downtime

This goes with the above. There is an underlying “of course you’ll work 100 hours a week” attitude which guarantees just shoddy performance and sloppy work. You’ll be judged on the result and not the process. No one can really tell how long it takes anyone to do some on the things you do. Play up that they were “laborious” or “very hard”. Practice your “I’ve got so much to do rant”. Paint dark circles under your eyes. Academia is horribly toxic work culture. STEM doubly so. 

No one by you will say enough. Don’t let your body do the talking and set rest times and rest activities (like running or reading the murderbot books for the Nth time). Whatever restores. Saturday is off limits. Sunday is ok for some light reading or prep for class. 

And if you need a quip to shut the all-nighter people up: “I don’t work 70 hour weeks, I did it right the first time.”

Tuesday, January 19, 2021

JWST doesn't have a problem? Happy to be proven wrong.

I wrote a blog post in 2020 (that's SO last year) about my worry for the JWST. It came partly out of a place of frustration of the past semester. 

The deadline was...not at a great time. But I would be hard pressed to point to a time this academic year that would have been good. Or even okay.


So maybe it partly was a Lucy-with-the-ball problem? Would the deadline be postponed at the last minute? During the semester from hell, we were kinda hoping for it? And there was precedence for this with the last Cycle 1 deadline. So anyone strapped for time/focus/mental energy would think twice about writing a proposal from scratch. 


So there was a timing problem and a hype problem. 


Timing - well 2020. Pandemic. Some light sedition. It's hard to think ahead 1.5 years into the future and about science. Nothing really that the JWST project could have done. Well maybe not announce at the beginning of my semester, expecting a result at the end. Still. Act of Deity kind of situation. 


Hype -- there was some clear group think for a sustained amount of time that sort of hinted at a HUGE oversubscription of JWST in the first call. Big instrument, big science, lots of time already allocated. It made sense. One could point to C28 for Hubble where the rate went to 1:12. Not great for science since it becomes hard to gamble with the time and do something truly new. 


But this (mostly over twitter but also back when talking to people in person. I miss people. And talking science in person), was the hype that built. Not all of it came from non-JWST folks too. So to me, looking from afar (like hinterland afar) started to conclude that this was going to be intense


Part of the hype was a byproduct of JWST project doing its due diligence. No these were not HST proposals, yes you need to do more work upfront. Get on this now now now


So time to adjust the recommendations.


More lead time? If there is nothing else to be done, announcing at the beginning of a typical academic semester and expecting proposals at the end...yeah the call could open earlier. Oh who are we kidding? We know how astronomers work...we've seen the plot. It could work in theory.


Set expectations? Also not something the JWST project was super keen on. What if they low or high balled their oversubscription prediction. Still. 4:1 is fine?


Split phase 1 and 2. Honestly I still don't understand why this decision to combine was made. 


Rolling or mid-cycle proposals. 1.5 years till science operation is a long time. More science might happen. I might actually sleep well for a night (more likely after jan 20 for some reason) and come up with a decent science case. Either call it early cycle 2 or simply adopt the mid-cycle proposals. 


So the news that I am now seeing (still via twitter, still from a lonely hill in the hinterland) is that it was a solid cycle. 30% women PIs (comparable to HST C28) and a nice spread of geography and new users. Whether it included people from all career stages is something I am curious about. Also curious if it was all R1s or if there was a decent spread too. Not all my questions have been answered yet. But I clearly was too pessimistic last time. Hurray. Happy to be wrong.