I have to say, an embedded system with 4MB flash and 4MB RAM does not sound small.
In my line of embedded systems work, the usual micro would have 64KB address space, partitioned into RAM and ROM. In such situations, I personally also abstain from C++ due to its extra memory/type overheads. Sure, you may count on having an excellent compiler to optimise out the extra code introduced from an inheritance structure, but most of the embedded (vendor specific) C++ compilers I've seen haven't been that good about it!
I still find C best for such tight memory situations.
Actually, the article appears to be mainly talking about a circuit assembly technique, using DNA, which can assemble at the nanoscale. Memory is just an example application, so it's not even really talking about memory!
Hmm... wonder if the/. readers even read the title of the articles...:-)
I used to be of the same opinion, until I tried to get my parents to backup their system. The problem -- they had no idea exactly what to backup and, if the harddrive did fail for some reason, they wouldn't know how to reinstall everything and set it all up again.
It's probably OK for most/. readers to just backup/restore "the crucial stuff". But the most convenient thing for most users would be to just backup "the whole computer" and restore the whole thing when needed.
Does anyone have any idea which TCP and/or telnet implementations are broken in this regard? More specifically, are the more "mainstream" implementations (Win32, Linux) broken?
Eh? I thought the Nagle algorithm was specially developed for telnet. RFC-1122 sections 4.2.2.14 and 4.2.3.4 describes the Nagle algorithm, specifically citing telnet as an example application that uses this TCP feature.
Or are you talking about a particular (broken) implementation that specifically disables this feature?
Nope -- they're not typically optoisolated. Such beasts may exist in industrial PCs and such like, but I reckon it's pretty unlikely that consumer PCs would have such ruggedisation.
But I would expect some kind of overvoltage or short circuit protection. At least a diode or two somewhere...
Has anyone managed to get Q3Atest to work, without having to be root, on a 2.2.x system?
I've installed the GLIDE 2.53 drivers from the 3DFX site, but the Device3Dfx drivers choke when compiling the kernel module, so I can't install the/dev/3dfx device.
I have to say, an embedded system with 4MB flash and 4MB RAM does not sound small.
In my line of embedded systems work, the usual micro would have 64KB address space, partitioned into RAM and ROM. In such situations, I personally also abstain from C++ due to its extra memory/type overheads. Sure, you may count on having an excellent compiler to optimise out the extra code introduced from an inheritance structure, but most of the embedded (vendor specific) C++ compilers I've seen haven't been that good about it!
I still find C best for such tight memory situations.
Actually, the article appears to be mainly talking about a circuit assembly technique, using DNA, which can assemble at the nanoscale. Memory is just an example application, so it's not even really talking about memory!
/. readers even read the title of the articles... :-)
Hmm... wonder if the
I used to be of the same opinion, until I tried to get my parents to backup their system. The problem -- they had no idea exactly what to backup and, if the harddrive did fail for some reason, they wouldn't know how to reinstall everything and set it all up again.
/. readers to just backup/restore "the crucial stuff". But the most convenient thing for most users would be to just backup "the whole computer" and restore the whole thing when needed.
It's probably OK for most
Does anyone have any idea which TCP and/or telnet implementations are broken in this regard? More specifically, are the more "mainstream" implementations (Win32, Linux) broken?
Eh? I thought the Nagle algorithm was specially developed for telnet. RFC-1122 sections 4.2.2.14 and 4.2.3.4 describes the Nagle algorithm, specifically citing telnet as an example application that uses this TCP feature.
Or are you talking about a particular (broken) implementation that specifically disables this feature?
Nope -- they're not typically optoisolated. Such beasts may exist in industrial PCs and such like, but I reckon it's pretty unlikely that consumer PCs would have such ruggedisation.
But I would expect some kind of overvoltage or short circuit protection. At least a diode or two somewhere...
/MC
Apache... you'd have to be running a web server of some kind to be affected at all.
Has anyone managed to get Q3Atest to work, without having to be root, on a 2.2.x system?
/dev/3dfx device.
I've installed the GLIDE 2.53 drivers from the 3DFX site, but the Device3Dfx drivers choke when compiling the kernel module, so I can't install the
I'm running a kernel 2.2.6 system, FWIW.