Knowing which projects/languages/frameworks to invest time into and which to skip (even if they produce useful subprojects) is a superpower these days.
replies(9):
I think it was quite successful, and I doubt they are shuttering it because they don't see the value in it, but because of overall lackluster company performance and the new CEO cutting costs/the workforce aggressively.
This happened recently with Scitkit-Learn Intelex, which was a drop-in replacement for some parts of sklearn that was a bit faster. One day, the Intel channel on Conda just stopped working (and I learned that Anaconda loses the will to live when a random channel you installed one package from is unavailable) and another organization took over Sklearn Intelex.
No communication could be found on Google connecting them to Intel (whose only news around the package was announcing the initial release a few years ago), you had to read the Git issue history to find people talking about the transfer.
I still have no idea what even happened to their Conda channel after the sudden disappearance. The complete lack of communication just left a bad taste in my mouth...
There is a lot of software that fits and optimizes for the former. Just be smart and selective about your choices, and avoid compromising.
"a theorized phenomenon by which the future life expectancy of some non-perishable things, like a technology or an idea, is proportional to their current age."
But I have no idea who’s actually paying the bills, behind the scenes.
A lot of normal users don’t feel these pain points or tolerate them by sidestepping the whole affair with containers or VM images. But if you’re in a position where these things have an impact it can be extremely motivating to seek out others who are willing to experiment with different ways of doing things.
There’s a tension there, but this is why it’s a skill — theres no simple rule. Fully open source community governed projects can be some of the most obviously good to ignore.
For a less controversial take, consider alpine's apk package manager. For a single-use container that runs one utility in an early dockerfile stage, apk can probably produce that image in 2-3 seconds, whereas for an apt-based container it's more like 30 seconds. That may not matter in the grand scheme of things or with layer caching or whatever, but sometimes it really does.
It requires a certain taste. There's a skill involved.
> Just be smart and selective about your choices, and avoid compromising.
This is a very "draw the rest of the owl" kind of statement
Is there a list, like there is for Google?
If you spit up chip design and fab, who would be interested in each? And is there enough x86 demand to keep the design side open? Windows on ARM is a thing, and data centers have been buying more from AMD than they used to.