herf

14 ms·

121.
by herf 5y ago
You can track handles within your process using GetProcessHandleCount. We have discovered some leaks in Windows APIs this way (which were fixed pretty fast).
122.
by herf 5y ago
We should remember that one of the problems with retail right now is wages that are too low, so low that many restaurants can't hire enough workers. These businesses are often paying for prime retail and then paying again for custome
123.
by herf 5y ago
When you look at the differences in speed between distros, often the main thing is which glibc is being used! Regular string ops can vary in speed by 2-5x based on which old glibc you have.
124.
by herf 5y ago
Did you install homebrew via a Rosetta Terminal? M1/ARM code is treated more strictly than Intel, so I guess all my command line stuff is Intel.
125.
by herf 6y ago
Upstreams were very limited, and HTTP requests send cookie data in each GET. For instance, if you had a 128kbps DSL upstream and each request was 2KB (loaded up with cookies), you're already limited to 8 requests/second. A cooki
126.
by herf 6y ago
Winamp was 1997. I remember there was one software decoder that was fast enough in 1995, so I plugged in my Pentium/100 laptop to decode music and pass it uncompressed to my desktop Pentium/90 (which was too slow) over 10base-T!
127.
by herf 6y ago
You can quite easily inspect the Great Suspender code (since it's just code in a folder), but it's not like Chrome gives you a diff when it updates. For months, nobody noticed, and there are still a lot more people who didn't
128.
by herf 6y ago
I think badly-behaved apps on the desktop can do this, or constant swap could as well. Even a poorly-configured redis could do constant writes.
129.
by herf 6y ago
At 5300MB/sec, it can exhaust its lifetime write limit in 31 hours? We need to start expressing write endurance as a multiple of write speed.
130.
by herf 6y ago
I think the big variable is how fast your organization is growing, and not so much the "kind" of organization. If you are in a place where everything's growing (more funding, more revenue), all your peers will get promoted wi
131.
by herf 6y ago
Shadow-mask CRTs were the best ever for low-res antialiased text. It looked great for the low resolutions of the time. (Even trinitron wasn't as good.) You were guaranteed a near-gaussian point in the output, and from a signal-processi
132.
by herf 6y ago
The day after this article (i.e. yesterday), some things are updating again: https://9to5google.com/2021/02/12/google-youtube-iphone-upda...
133.
by herf 6y ago
Recent notes are all about HEIF (~= iOS HEIC). It seems alarming to me that so many devices (Canon and iPhone at least) are putting out patent-encumbered HEIF/HEIC files while the available decoders are so incredibly slow. Last I loo
134.
by herf 6y ago
In Thunderbird, OAuth2 login is still broken. The login page prompts for email again and again, never makes it to the password.
135.
by herf 6y ago
Looks like they treat "opt in" data collection the same way they treat silent collection that's turned on by default? This seems like a bad incentive, because there should be a benefit to being careful about opt-in.
136.
by herf 6y ago
This is sort of backwards from an economies-of-scale perspective. In the normal world, shipping your software internationally is something of a cost (like filing VAT in all the countries). And even though Stripe has made it way easier to do
137.
by herf 6y ago
We still have awful authentication options on the LAN, so going through the cloud fixes this. One advantage I like is you can put IoT things on a segregated network.
138.
by herf 6y ago
Anandtech is reporting 62GB/sec memcpy, and the way-faster Geekbench results are the memory-bound ones (e.g., Gaussian blur is 4x faster!) So I think the speedups like these are largely due to memory architecture.
139.
by herf 6y ago
I am wondering if M1 is winning many of these due to main memory latency? I think Geekbench 4 reported memory latency, but it's not in v5. Does anyone know of a benchmark of this directly?
140.
by herf 6y ago
23andme doesn't report missense variants, so it doesn't report this.
141.
by herf 6y ago
No, ozone only happens below 210nm, which is why narrowband 222nm is good.
142.
by herf 6y ago
This is well argued, and I think the big change there was the early-2000's DRM ideas: that a hardware platform like the iPod could protect the IP used on it. Initially this was seen as important for music, even though the restrictions
143.
by herf 6y ago
I've been disappointed with CPU/battery usage recently. When adding a lot of photos, I have to plug in my laptop--otherwise the battery dies. When uploading a 100MB file on a gen8 Thinkpad: - Dropbox uses 6 minutes of CPU time, ba
144.
by herf 6y ago
Yes agreed, they say "structural color" (film interference) only works for blues and greens, and you need pigments for oranges and reds. Title should be changed.
145.
by herf 6y ago
There is a problem with "cheap" monitors and DDC/CI: some of them use EEPROMs to store brightness settings, and this limits you to about 100,000 writes. Worrying about this is the main reason we don't ship DDC/CI wi
146.
by herf 6y ago
App Review should be more transparent and more public. For instance, even without naming an app, the reasons for rejection by category could be public. In the bigger discussion, we don't know if "difficult" cases like the one
147.
by herf 6y ago
I think shippers like Amazon pay relatively high rates to USPS, so they are higher on the priority list. It's the heavily-discounted "slow" shipping like medicine that is having trouble right now.
148.
by herf 6y ago
I keep trying to understand if these ideas will be useful for epidemiology. Right now I think there is a long way to go for multivariate statistics. It seems that differential privacy can handle one column of data, with some categorical fil
149.
by herf 6y ago
This is a fine bit of research, but exclusivity should be the first question--the high-order bit. There has been essentially no market discovery of prices because of exclusivity. Since 1992, there has been "free" distribution onli
150.
by herf 6y ago
It's important when doing these kinds of analyses to consider the effects of DST. The first graph in the article depicts April 2, 2019 sleep times vs April 7, 2020. But April 2 was just a day after the time change (Dec 31 2019), while

More ›