Last week on HN, a non-student was complaining about having to increase DrRacket's memory limit 3 times while they were playing with it, so I pointed out that DrRacket was designed for new students, and suggested that maybe that memory limit was a good thing for new students.

One of frequent complaints, from generation to generation, seems to be "kids these days got it too easy".  Which, in programming, is not necessarily grumpy, but concern that a lot of learning opportunity that comes from working with resource constraints is being missed. Personally, I think there's a balance, and there's also learning opportunity missed when you don't have lots of resources.  Ideally, a person gets both kinds of experiences.

As a C and C++ programmer who was then an early Java promoter, I was a bit concerned about that, at the time.  I figured we'd probably move to Java, and all the students already had use of powerful multiprocessor workstations.  That was one of the attractions of then playing with programming the Pilot PDA ("";), and I promoted Pilot programming to other students specifically for the reason of learning to develop with tight resource constraints.

Other Racket relevance: the approach to fitting the toy "route planner" into a few KB was to use two little DSLs, with a Lisp as code generator to get around the limitations of macro preprocessing in C.  Between that, the crazy DSLs I made as sets of cpp macros for my compiler (C++) and robot (C) assignments, and an awful concurrent hierarchical state machines language that took way too much effort to implement in Java, I suppose it's not a surprise I later decided to move to Scheme/Lisp for my main research tools.

Also, copious computing resources becoming available to lots of people became a concern to some researchers, who no longer felt as privileged as before.  At the start of the dotcom boom, one of my grad school advisors was already spending most of their time on startups (and there was some truth to the joke about advisor only wanting MS+IPO students).  They called an off-site retreat for our group, where a big part of the case was that the university research lab no longer had special advantages like supercomputers that other people didn't have.  (And I'd previously been horrified to see the storage array cabinet from a Connection Machine being used as a barkeeper counter, for the lab's posh sponsor events.).

Today, it's true: I have my own deep neural networks supercomputer in my living room, for a few hundred dollars, and it's just an ordinary consumer GPU like children have in their gaming PCs/consoles.  Which makes for "lots of resources" learning opportunities, complementing the "not enough resources (until you figure it out)" learning opportunities.

You received this message because you are subscribed to the Google Groups "Racket 
Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
To view this discussion on the web visit

Reply via email to