Re: GETF Speedup?

> On Oct 15, 2025, at 1:51 PM, Tim Bradshaw <[email protected]> wrote:
> 
> If I had a pound for every time someone explained to me that machines I used were user-microcodable I would have many pounds.  If I had a pound for every time someone showed me an example where that really helped things significantly for any length of time I would have the same number of pounds I have now.  It didn't even make Lisp fast, at least compared to expending the same resource on compiler development would have.  Hennessy and Patterson were right, it turns out.  I am sure there is some famous example where user-written microcode made all the difference.

Yes, RISC v CISC. And they pushed caching. But that doesn’t help with GUPS (if anything, caching hurts). If all you are interested in is making your existing imperative codebase faster, then yes, you are probably better off investing in “sufficiently smart compilers”. If you are instead interested in alternative architecture and algorithms to avoid von Neumann bottlenecks, compilers won’t help as much as more control over the hardware. (That’s forced me to work with FPGAs, but I’m still building microprogrammed processors in them). 

> 
> The keyboard point is better: the Sun type ?3? keyboard was a lovely thing (I never had time for the crust of cokebottle keys that surrounded the spacecadet-derived keyboards).  But you can get modern keyboards with good keyswitches and good layouts: I just am not willing to trade a laptop for something with a separate keyboard.
> 
> It's like saying a sidevalve Aston Martin is a more lovely car to look at and drive than anything modern.  It is.  But I'd still like to drive an electric car with brakes that work and seatbelts, on the whole.
> 
> I'll go further: LW, today, is an environment I find more pleasant to use than I ever did Genera.  We're reminiscing about a partly-imaginary lost past which wasn't, in fact, better than the present.

I’m not arguing against LW, and I have a paid version, as IMO it’s the best of what’s supported today. I am arguing that there were innovations from the past that, perhaps, were too casually discarded because they weren’t a feature of the commodity processors, or seemed hard to implement/support. Most of these kinds of arguments have already been made in the Unix Hater’s handbook (you can read it here: https://web.mit.edu/~simsong/www/ugh.pdf), so I hesitate to repeat them here. 
> 
> I know: heresy.
> 
> --tim
> 
>> On 14 Oct 2025, at 23:32, Bradford Miller (as bradford dot w dot miller at gmail dot com) <[email protected]> wrote:
>> 
>> Not everything. Vintage machines (not just LispMs) allowed user microprogramming, for example. And the Kulp keyboards were definitely better than anything commercially available today. And don’t get me started on everything we’ve had to give up in the name of “security”… 
>> 
>> Sure, even SBCL would blow an XL-1200 away on Gabriel Benchmarks. But there’s a large difference between desiderata for development v. deployment.
>> 
>>> On Oct 14, 2025, at 4:55 PM, Tim Bradshaw <[email protected]> wrote:
>>> 
>>> On 14 Oct 2025, at 20:53, Bradford Miller <[email protected]> wrote:
>>>> 
>>>> I never said there was. I’m just disappointed (?) that many such LispM ideas haven’t been preserved in current platforms.
>>>> 
>>> 
>>> I miss LispMs in the same way I miss vintage sports cars: they were absurdly cool and fun ... but realistically everything (LW, say) is better today.
>>> 
>>> (And yes, I used LispMs, and yes I've driven cars made a century ago.)
>>> 
>>> --tim
>> 
>> 
>> _______________________________________________
>> Lisp Hug - the mailing list for LispWorks users
>> [email protected]
>> http://www.lispworks.com/support/lisp-hug.html