I wonder if this is because Microsoft Research is no longer interested in funding his research, or if for whatever reason SPJ is just seeking greener pastures.
Either way, it's a blow to MSR - SPJ is such a luminary in the world of Haskell/PLT. Best wishes for whatever he winds up doing next; no doubt it'll be great.
I've shipped stuff written in Haskell almost 2 times... but actually one time.
It's not that bad. Honestly the only reason we shipped it was because the prototype I did in it worked too well.
I had a backup almost ready in time written in C++ in case it got too tricky for others to maintain, and it was a simple component in a larger software stack.
I would be careful making such generalizations lest you give this vocal minority of haskellers more reason to believe that non-haskellers are luddites (edit: can you use the word "luddite" to refer to people who don't like theory? Probably not but pretend that's what it means).
I've used Haskell for a few personal projects (GHC 6.8-7.0 timeframe). It is great but I stand by my statement that the community's obsession with complexity (even if it's simpler from a category or maths sense) is the issue. Haskell's community is almost *obsessed* with increasing cognitive load IMO.
I want to make things not argue about the best, or most technically correct, way to make them.
> It is great but I stand by my statement that the community's obsession with complexity (even if it's simpler from a category or maths sense) is the issue. Haskell's community is almost obsessed with increasing cognitive load IMO.
Personally, I consider it more complex to have to think about what side effects every piece of code has and what order things are evaluated in. The fact that I don't have to worry about these things in Haskell is a breath of fresh air for me. Also, not having to deal with the complexity of OOP when an ordinary function will do.
I think it's a misconception that Haskell is more complex than whatever popular OOP language people are using. It just looks more complex because it's less familiar (which I think is a problem with what we teach students in school).
> I want to make things not argue about the best, or most technically correct, way to make them.
Well, not all programming languages are for all people, and there are people who prefer the theoretical approach. Calling that an obsession with complexity is, I think, unfair, or at least inaccurate.
"If your confusion leads you in the right direction, the results can be uncommonly rewarding" --Haruki Murakami, Hard-Boiled Wonderland and the End of the World (as translated to English)
Eh, I don't think it's any different than the general trend of people moving around. 23 years at a company would indicate that it's been a pretty solid place to work.
At the same time, leaving after 23 years for a reason other than retirement suggests something changed.
I have no insight into the company as you do, but that is the first impression I’d get from the circumstances. People don’t stay at a company for 23 years only to jump ship for the same reasons early-career people jump around.
He's been at Microsoft for 23 years and it's basically the only company he's ever worked at. He's also over 60. I can't blame him for wanting to try something else with the tail end of his career.
SPJ had a great deal of latitude at MSR Cambridge, he could probably work on almost whatever he wanted. The problem is MSFT is rife with bureaucracy and it is a constant fight to get anything done. MSFT in general and MSR Cambridge specifically has lost a lot of their best people over the last 10 years. AFAIK it is expected to get worse under Christopher Bishop. With the recent win in Excel Lambda it is as probably a good time as any to get out.
The fact that there is more discussion here vs the Haskell forum is weird. I've not used Haskell myself but aspire to use it someday. I wish him best of luck!
Anecdote time: I attended a Haskell event at MSR some fifteen years ago, and one of my overwhelming memories is how approachable and friendly SPJ is. I had some pretty stupid questions around the topic he was presenting(STM), and he answered them with great vigour and clarity afterwards.
This does feel like a significant loss to MSR, and a massive gain for wherever he ends up. Fingers crossed it is a teaching role, as I'd like to think we'd all benefit from turning a few more SPJs loose on the world.
There is a recent MSR podcast episode with SPJ and Andy Gordon that is definitely worth a listen¹. It dips in to their histories, and also their work on Excel.
Edit: My other memory was the sheer level of professionalism in the support staff at MSR Cambridge, I simply can't recall anywhere I've been where the support staff have been as helpful and courteous as they were there.
26 comments
[ 2.4 ms ] story [ 66.4 ms ] threadEither way, it's a blow to MSR - SPJ is such a luminary in the world of Haskell/PLT. Best wishes for whatever he winds up doing next; no doubt it'll be great.
This is what got me with Haskell. Any question is derailed into theory to the point where getting anything done is difficult.
It's not that bad. Honestly the only reason we shipped it was because the prototype I did in it worked too well.
I had a backup almost ready in time written in C++ in case it got too tricky for others to maintain, and it was a simple component in a larger software stack.
I want to make things not argue about the best, or most technically correct, way to make them.
Personally, I consider it more complex to have to think about what side effects every piece of code has and what order things are evaluated in. The fact that I don't have to worry about these things in Haskell is a breath of fresh air for me. Also, not having to deal with the complexity of OOP when an ordinary function will do.
I think it's a misconception that Haskell is more complex than whatever popular OOP language people are using. It just looks more complex because it's less familiar (which I think is a problem with what we teach students in school).
This is a problem I have at work. Everyone here thinks unfamiliar and hard / complex are synonyms. They aren't!
Well, not all programming languages are for all people, and there are people who prefer the theoretical approach. Calling that an obsession with complexity is, I think, unfair, or at least inaccurate.
Eh, I don't think it's any different than the general trend of people moving around. 23 years at a company would indicate that it's been a pretty solid place to work.
I have no insight into the company as you do, but that is the first impression I’d get from the circumstances. People don’t stay at a company for 23 years only to jump ship for the same reasons early-career people jump around.
This does feel like a significant loss to MSR, and a massive gain for wherever he ends up. Fingers crossed it is a teaching role, as I'd like to think we'd all benefit from turning a few more SPJs loose on the world.
There is a recent MSR podcast episode with SPJ and Andy Gordon that is definitely worth a listen¹. It dips in to their histories, and also their work on Excel.
¹ https://www.microsoft.com/en-us/research/podcast/advancing-e...
Edit: My other memory was the sheer level of professionalism in the support staff at MSR Cambridge, I simply can't recall anywhere I've been where the support staff have been as helpful and courteous as they were there.