dfox

74 ms·

841.
by dfox 7y ago
The F1 engine is interesting from this PoV because the thing as it is documented does not work (and even the documented design is full of weird design decisions that nobody today knows the reason for) and it worked by combination of luck an
842.
by dfox 7y ago
The reason for warnings on 11 was that the landing was done in what was essentially an unsupported configuration. Still the “kernel” behaved as designed, didn't run low priority task (that was not supposed to be running in the first pl
843.
by dfox 7y ago
Assuming reasonable charger design (ie. at least some overcurrent protection) the internal ESD protection of USB-C port in the device that only ever expects 5V (think SRV05-4) should be more than enough to safely trip the overcurrent protec
844.
by dfox 7y ago
It suffered from simple CPU overload, in order to have priority inversion you need to have preemptive multitasking and synchronization primitives. The scheduler on AGC was essentially cooperative. Side note: that is partly what the 90'
845.
by dfox 7y ago
I assume that the issue with processor groups on windows is that something somewhere is hardcoded to represent some kind of per-CPU state as bits in machine word. Having some kind of hypervisor as an additional layer will not help with tha
846.
by dfox 7y ago
I had firsthand experience with this otherwise useful feature masking an obvious bug. In one codebase there was function along the lines of (for calling into MSSQL with it's "peculiar" stored procedures): def call_store
847.
by dfox 7y ago
It isn't. Because when they started, there wasn't any "outer" platform the speak of. The platform underlying SAP is essentially an way to have mainframe-like development environment (updates of almost anything without re
848.
by dfox 7y ago
If you think, that SAP's UI is inconsistent then you haven't seen the above mentioned EBS ;) Input boxes behaving as buttons, tabs either changing behavior of UI components that are clearly drawn as being outside of the tab, or ou
849.
by dfox 7y ago
I assume that the sound refers to fact that about half of Windows SAPgui components are implemented by embedding MSHTML, which by default makes "click" sound for every onload event.
850.
by dfox 7y ago
one has to differentiate between SAP the product and SAP as implemented by random big-co. The parts that are user actively confusing and user hostile are usually result of customization done during the implementation (and motivated either b
851.
by dfox 7y ago
The reason why people mostly meand second pre-image when saying unqualified "pre-image" is that probably any imaginable method of reversing a hash function (given sufficiently long input to the hash) will with overwhelming probabi
852.
by dfox 7y ago
There is no technicqal reason for using unique IVs in this case. But it does not hurt and looks better in security audits. Edit: there is also an line of reasoning that you should not needlessly expose cryptographic primitives to "non-
853.
by dfox 7y ago
True HSM with active self destruct needs to be constantly powered. On the other hand for many if not most applications, typical secure smart-card is is completely sufficient (and in fact typical POS card terminal stores most of it's lo
854.
by dfox 7y ago
I somewhat believe that the sole reason for existence of PostThreadMessage and friends is to make COM cross-apartment calls work for threads that do not own any windows. You can probably devise other uses for that, but such uses still somew
855.
by dfox 7y ago
For me this is the only relevant reason why you don't want to do multithreading or alternatively why you want to structure your multithreaded application as mostly independent “processes” that communicate by means of message queues. I
856.
by dfox 7y ago
By the way this is how threads work on Linux to this day. LinuxThreads were exactly this and NPTL is mostly about doing magic with RT signals and shadowing libc symbols such that stuff that changes global process state works reliably. Along
857.
by dfox 7y ago
The message queue is user-space abstraction relevant to the GUI subsystem and not something that is inherent to Win32 threading model. And in fact its implicit existence is source of completely unique Win32 class of hard to debug threading
858.
by dfox 7y ago
I'm aware of that. But that is also partly my point, before ribbon there was 'File->Import...' which did essentially the same thing but at least for me (and many users I work with) was more discoverable. Another thing is t
859.
by dfox 7y ago
Excel actually mostly ignores the extension and relies on heuristics to detect the file type on load. Many web applications used to implement "XLS export" by writing out CSV file with .xls extension. This magic detection of file t
860.
by dfox 7y ago
The question is how practical that mode of operation is on real network, or in other words whether it actually makes sense to run DOCSIS on network that can reliably support that mode (think various xDSL technologies that can push >1Gbps
861.
by dfox 7y ago
The reason why it is easier to filter by wavelength is that almost any optical path reflects significant amount of light back, primarily from various discontinuities along the fiber (connectors, splices, and also the actual receiver on the
862.
by dfox 7y ago
Forth is more low level, ie. there is no object memory/GC at what one would describe as implementation level. On the other hand all the Lisp Machines essentially are hardware implementations of something vaguely Forth-like with some le
863.
by dfox 7y ago
That is how the normal POTS phone works and there is some expectation that physical SIP VoIP phone would work in same manner. For the VoIP phone the usual implementation is that the hardware SIP phones have configurable regular expression t
864.
by dfox 7y ago
H.323 is the internet protocol designed by phone people, SIP is prime example of second systems effect caused by internet people trying to make it “more internet-ish”. The reason why H.323 is “unspeakable horror” to internet people is mostl
865.
by dfox 7y ago
If I think about it I probably can pinpoint the reason for the soldered cable looms: not only in SSSR but across the whole eastern block in general there simply weren't any reasonably cheap and reliable board to board or board to cable
866.
by dfox 7y ago
Assuming this thing was designed in early 80's (which it probably was given the LQ470-like DIL 7segment LED displays) there certainly was soviet made NMOS ASIC that implemented most of the functionality in single package. But similarly
867.
by dfox 7y ago
Wrt. to flatpaks it is somewhat interesting to note, that very often when there is some kind of ASIC (think alarm clock or calculator or even “PDP-11 on a chip”) in 80's soviet consumer electronics it is packaged in metal/ceramic
868.
by dfox 7y ago
Surprisingly large amount of soviet era consumer electronics are built quite similarly and use what in essence are military/space grade parts (and meticulously made cable looms). I assume that this is simply caused by using whatever wa
869.
by dfox 7y ago
The problem is that on i386 the TLB entries are keyed only by linear address and not by page table that the entry came from and writing into the page table register (even rewriting the same value) is documented as triggering TLB flush. On t
870.
by dfox 7y ago
IIRC enough of the KBC is still present that this might actually work, another question is whether modern UEFI firmware actually handles the SW-induced reboot. As for entering and exiting protected mode: that is in fact surprisingly fast op

More ›