nimrody

> You can verify that your implementation is capable of performing lock-free atomic operations on double-machine-word primitives with assert(data_ptr.is_lock_free()).

This may be true when compiling for a 32bit process and running on a 64 bit machine, but is not true for the more common 64 bit program on x86_64 (since pointers are 64 bit.)

My ipad 2 is certainly much slower tan it used to be. I believe the reason is the new iOS versions. Even switching between apps feels annoying.

Using Fedora 25 building Android 7 for Nexus 5X is relatively simple. The only caveat is that to really take advantage of the hardware you need vendor blobs which you can get by running the scripts at https://github.com/anestisb/android-prepare-vendor

Beautiful work!

It's nice to see something that was designed with maintainability in mind. Designed to be disassembled, repaired and re-assembled later. Impressive engineering.

So different from most consumer products sold today which never use screws and are not designed for repairing. If it breaks down you're expected to buy a new one...

Even the method cache seems expensive -- compared to an indirect function call (using a virtual method table). Can clang replace msgSend with direct calls when the destination class is known at compilation time? (perhaps with a guard to verify that the object class is as expected)

Then perhaps Andreas Gal's "Chrome Won" assertion wasn't entirely accurate. After all WebAssembly is something that was derived from Mozilla's asm.js.

Unlikely given any non-negligible chunk size and accounting for the randomization effect of compression (image and video).

I think deduplication buys them most when people upload common assets - videos downloaded from the net, ebooks, ISO images, etc.

My guess is that it's Gnuplot. At least gnuplot has an SVG driver: http://gnuplot.sourceforge.net/demo_svg_4.6/

At least part of their content delivery network is based on their own hardware/software (see https://openconnect.netflix.com/en/ and the PDF referenced there). It's called "Open Connect Appliance".

They use AWS for video transcoding) (see http://techblog.netflix.com/2015/12/high-quality-video-encod...) and probably other activities (analytics, business)

Right. Something about hardware and OS staying stable long enough so that people could focus on creating new interesting software -- instead of just trying to keep up with ever changing APIs and constantly searching StackOverflow for ready made solutions.

Sometimes severe constraints drive the really beautiful solutions.

Would be interesting to implement a Linux compatibility layer on top of ReactOS -- just to get the benefit of using binary drivers supplied by manufacturers for Windows. Perhaps Wayland could run on top of ReactOS and benefit from native DirectX drivers?

How do you manage multiple environments (staging / production) with these type of tools? How do you avoid making changes in the wrong environment?

A single Postgres machine holds all data? So you scale the database by switching to a stronger machine? Or do you plan to shard the data somehow?

Do you have any queries iterating over the entire table?

Why would the browser need to reload all resources after a POST request?

Presumably, only the main (HTML) page should be refreshed and other linked items (CSS/JS/media) with high max-age (and no 'must-revalidate' cache control header) should use the previously cached content.

There's a Fiddler version for Linux (based on mono) which, in my experience, works quite well. Easier than starting up a windows VM.

The killer feature of Fiddler, for me, is the timeline display which we use to optimise network heavy apps.

Would be useful if AVRO files were supported. This was the data can also be imported into Redshift if needed (Redshift does support Avro).

Other formats are schema-less (JSON,CSV, etc.) or not supported by Redshift (ORC, Parquet). Perhaps less efficient for some queries (AVRO is not a columnar format) but still useful.

How much battery life do you get from this configuration? macOS is optimized for power. FreeBSD not so much, I believe.

Not that I'm an Oracle fan, but they do continue to develop MySQL. Both in terms of features and performance.

Oracle still invests heavily in the development of the JVM (graal and truffle aim to advance the state of art in JIT compilers)

So it's definitely not all bad (virtualbox is also used by many but unlike MySQL I don't follow the development)

Today even smartphones are ARM64. They are powerful, yet power limited.

> There are still requests sent, there is still network latency. [In the case of small JS/css/icon files, the latency can be as much as the transfer time].

What? When the "Expires" header indicates the item is still valid the browser should use the cached object without making any requests.

Some servers (Rails for example) just serve assets with "Expires" set to some distant future time. When they want to expire these files, they simply change the name (hence the long numeric suffix appended to js/css/etc. files)

Pascal and Ada don't usually use their arrays as matrices in mathematical expressions. Julia is targeted at people using Matlab.

How do you multiply two matrices that use different indexing? Ignore the indexing differences?

It's also problematic when different libraries use different conventions (0-based vs 1-based). Can you use the output of one such library to feed another?

Dropbox say they can losslessly compress JPEG files (~20% saving):

https://blogs.dropbox.com/tech/2016/07/lepton-image-compress...

Do mobile networks (cellular networks) these day commonly allow anything but TCP? (or even HTTP? Most firewalls block anything that is not HTTP/HTTPS)

No idea. There's an app that's supposed to test for doze and it says "hw sensors: ok" (or similar) but "config: off". Like it's configurable somewhere but wasn't enabled by Xiaomi.

My guess is that Xiaomi already has a similar feature (perhaps less sophisticated) called "Autostart". Most app aren't allowed by default to wake up in the background and kill your battery. Only some (e.g., whatsapp) are whitelisted and, of course, you may whitelist additional apps yourself.

Overall, I'm happy with the battery usage of my Mi3. I prefer the Xiaomi Apple-like UI over Google's standard -- but that may be just a habit...

My Xiaomi MI3 (international version) is almost two years old and now got automatically upgraded to Android 6.0.

(No doze support and it revoked my root access but still nice!).

Priceless but also very expensive :) I wonder if someone researched the cost of raising a child all the way to adulthood.

How does ToroDB handles sharding across multiple instances?

No. Only TLBs are flushed (and probably only partially). TLBs are used to associated virtual addresses with physical addresses and memory maps are different per processes.

(That's one reason why it's beneficial to schedule a process on the same CPU if possible - the data is still in the cache)

With Genymotion you can run multiple emulators on a single machine. The Intel tools you mention (Intel HAXM for fast virtualisation) seems to work only for a single instance. The second instance does not benefit from fast virtualisation and is painfully slow.

Don't know whether this problem has been addressed by this 2.0 release.

It depends on whether you have many small files or a small number of large files. In the former case, one file approach (HDF5) makes sense. In the latter case, you don't have a problem with i-node consumption and you gain the ability to easily access only the data you need without bringing the entire data-set into memory.

I kind of the like the approach suggested above - if you have many small files store them in a zip archive and use some library to access the data directly.