Limitatations can enable innovation.
I also have a T40p, from 2003-ish, with the same 1400x1050 display. This is an even nicer computer to use in many ways - the keyboard in particular is even better than that in my T60 - but the difference in performance between 2003 and 2006 is striking. It's almost as much as the difference between 2006 and now. The T6x used Intel Core CPUs, but the T4x was still Pentium M days - single core and quite substantially slower. The T4x still has a PATA drive, so it's harder to put in a modern SSD. It's limited to 2G RAM for reasons I can't remember. It's also much less robust - the T60 series has a completely rigid chassis (and my T60 still looks like new) but the earlier T40 would bend if picked up from the front, which would cause both the case and the tracks on the motherboard to crack (the USB sockets stop working first, then everything else).
The T40p is a lovely, lovely computer - in some ways the best one I own - but it's for occasional document-writing only, not really for development any more.
Maybe I should look for a fancier one, if only to pick up the motherboard and CPU from. I imagine they're quite similar in terms of performance, but the extra compatibility would be nice.
I have a T500 Core2Duo T5600 w 4GB RAM sitting in the cupboard. I pulled it out and put an SSD in it and installed Lubuntu, works fine but I have some software licenses I’ve paid for that are only available on MacOS & Windows.
Do you think it would run Windows 10 ok? I’d need to buy a license, but at this stage that might be cheaper than a new / secondhand laptop.
Memory hungry apps like Chrome slow down pretty quick, but the OS itself is snappy.
This is a very good point. I think the user experience, on average, would be quite a bit better if developers weren’t so prone to chasing glitzy new hardware that can “mask” performance problems.
On the other hand, developer’s productivity matters, so I’m not really sure what the right balance is here...
Also, I find that new developers think they would benefit greatly using the fastest machines to learn to program, but then never develop a good sense of algorithmic complexity as everything just screams. On slower machines you can have the opportunity to feel the difference between O(1), O(log n), O(n), O(n^2), etc.