jedbrown

42 ms·

361.
by jedbrown 14y ago
One of my biggest complaints about the current state of patents is that it reverses the process of innovation. I have not heard of anyone reading existing patents to get new ideas for software. Instead, we invariably invent first , usually
362.
by jedbrown 14y ago
I can think of very few algorithms that don't have sub-algorithms. Isn't any composition of existing algorithms also an algorithm? I can't think of any restriction that algorithm (singular) makes relative to the plural.
363.
by jedbrown 14y ago
Bah, I can't think of a case where the user interface of ftp is so precious it can't be replaced with rsync or sshfs.
364.
by jedbrown 14y ago
Writing good libraries seems to take a reasonable amount of experience. Of course if you have been doing a significant amount of software as a student, you'll have a head start. My skills improved a great deal when I transitioned from one-o
365.
by jedbrown 14y ago
If you have top quality work, where are you going to publish it?
366.
by jedbrown 14y ago
I have yet to meet someone with decent software development skills and a degree in a scientific/engineering field that had any trouble finding good work. I have seen a lot of unfilled postdoc and higher positions due to inability to find co
367.
by jedbrown 14y ago
The viability of separate processes is highly dependent on the problem domain and algorithm. Threads have their place.
368.
by jedbrown 14y ago
Several comments in the thread point that out, but he replied: "It seems the cartridges are all full size. So it's really cheaper to buy a new printer."
369.
by jedbrown 14y ago
This only helps with angular momentum, not with linear momentum.
370.
by jedbrown 14y ago
Yes, all the passwords in LinkedIn's massive leak were unsalted.
371.
by jedbrown 14y ago
It looks like that is what LinkedIn claims to have done. http://blog.linkedin.com/2012/06/06/linkedin-member-password...
372.
by jedbrown 14y ago
You talk as though 2003 was prehistoric.
373.
by jedbrown 14y ago
Unsalted was not "state of the art" in the 1970s. The database would be far more secure if it had used a weaker hash with salt. To migrate, Unix-like systems generally support database migration through a modular format (see "man pam_unix"
374.
by jedbrown 14y ago
PHK is not concerned with preventing precomputed (rainbow table) attacks. A salt is fine for that. He wants to make it harder to create special-purpose hardware for brute-forcing. Combining many different types of hashes is only intended to
375.
by jedbrown 14y ago
Anyone who has put sled dogs in harness know that they are extremely eager and don't need coaxing. They also go far further than 100km speeds well beyond what people can do. The dogs lose their underfur in summer and can cover long distance
376.
by jedbrown 14y ago
The 100 mile (160.9 km) human record is 11:28 (on par with the camel 144km). The six day human record is 1038km (average sped of 7.2km/h). http://www.iau-ultramarathon.org/images/file/World_Best_Perf... Another data point is that sled dog
377.
by jedbrown 14y ago
Depends on data dependencies and the number of registers, but yes, L1 latency is rarely the bottleneck (more likely to be load/store throughput, in-register shuffles, or arithmetic dependencies for such tight kernels).
378.
by jedbrown 14y ago
Well, when people say 50% overhead (the lower bound), they mean that the cost to the grant is `1.5*direct_cost`, not that half the grant goes to indirect costs. Anyway, research is the primary motivation for purchasing journal subscriptions
379.
by jedbrown 14y ago
I doubt the 10-20% number as well, but note that that "site-wide access" is essentially paid for out of research funds. It's called "overhead" and ranges from about half of research salary to double or more, depending on the institution. Th
380.
by jedbrown 14y ago
Managing lists of publications is time consuming, especially when different publications should be selected in different contexts (e.g. choose 10 relevant publications for each grant proposal, or sort them according to different criteria).
381.
by jedbrown 15y ago
> Last week I spent 4 hours trying to find what turned out to be a missing = because C++ will not complain, but will behave very differently, when = is confused with ==. Uh, all reasonable compilers warn about ambiguous use of = as a tr
382.
by jedbrown 15y ago
If you make a huge allocation (many pages), isn't the Go runtime very likely to call malloc()? For large allocations, malloc() is going to get you a bunch of fresh pages and you will generally get the address of the start of a page. The off
383.
by jedbrown 15y ago
That would seem to also present a DOS vector (even on 64 bit) if a user can get the program to store data (of any type, e.g. char or floating point) that happen to be binary-equivalent to pointers to large allocations.
384.
by jedbrown 15y ago
However, it does not address the base issue, which is that Go uses a conservative garbage collector, and more values look like pointers in a 32-bit world. The only real fix would be to improve the garbage collector's understanding of which
385.
by jedbrown 15y ago
We won't know whether Larrabee was a screw-up until we see how successful Knight's Corner is. A lot of groups are currently running prototypes under NDA. It's a programming model debate as much as anything, the hardware capability is availa
386.
by jedbrown 15y ago
I write a lot of code that runs in funny environments like the Compute Node Kernel running on Blue Gene systems, with offset-mapped memory (no TLB) and no over-subscription. It's fantastic for reproducible performance, but it's a specialize
387.
by jedbrown 15y ago
You need an event-driven system because the number of "tasks" rarely matches the number of physical cores available. One tried and true, fault-tolerant event-driven interface is provided by the kernel: processes (and threads). It has interf
388.
by jedbrown 15y ago
"Any device maker who doesn’t contribute financially to the testing won’t be added to the new updated list of approved electronics devices on planes." Really? Your suggestion is that someone checks specific models against a special approved
389.
by jedbrown 15y ago
Or use git citool or an emacs mode like egg or magit (other editors have similar) that let you conveniently stage by hunk or line.
390.
by jedbrown 15y ago
The conversion is optimized out by gcc-4.6 at all positive optimization levels, the addition is optimized at -O2 and above. But the slowness is caused by the algorithms being different, with one producing denormals while the other stays rea

More ›