Microsoft's effort is about as competent as it can be in that it's seamless and *most* things sort of work. However *most* is not the same as *all* which is what people expect when they run Windows. And it doesn't account for the degraded performance and battery life when something runs under emulation. I feel sorry for the poor bastards who unwittingly bought a Windows on ARM device and then discover this. And I don't see the situation being any different for an open source ARM emulator. It'll do its best in the circumstances but it won't run as well as native instructions.

But why should a developer bother if they have this emulator thing there? One choice involves a bunch of work (having to build, package, test and deploy yet another build), the other choice means doing nothing. I honestly think the better option would be for the build target to be architecture agnostic, a universal target and then the game runs anywhere there is a compiler.

Even if the emulator forwards calls to system libs when it can (e.g. graphics), or caches, precompiles code it will still find itself bogged down in big chunks of x86 code which are is going to hog CPU & battery trying to execute. I would have thought that these negate the reason for using ARM in the first place. We've seen even Microsoft's own ARM laptops chug trying to emulate x86 software and the results are all over the map. In some cases games work acceptably, and in other cases abysmally or not at all. I certainly don't see it as the future of gaming to be doing this, even if it means some titles can be run on platforms they weren't originally intended for.

Instead Valve should be encouraging games to use some kind of universal binary format, e.g. based on LLVM bitcode which can be compiled against the target platform / runtime prior to download. i.e. when you download to your ARM device you get an ARM binary, when you download to x86 you get an x86 binary. These could compiled by the server rather than expecting every device incurring the overhead of making sense of a Windows x86 binary dumped on their laps.

Game devs chip in and fund a font that they can all use royalty free. Even better, open source it and shit all over the profit model of the company holding them for ransom.

Ah yes, me posting a comment on a niche website makes me an influencer. Jesus fucking Christ on a clown bike.

I'm fairly certain I am not the cause of this research.

So Swiss scientists were giving mice tramp stamps?

I don't consider myself a DB guy but I've used postgres to model graphs with potentially millions of nodes, each with json fields, recursive queries and a bunch of other very complex code either written directly or through hibernate / spring. Postgres has handled it fine. It seems like a very capable database and any issue with performance issue I find tends to be bad code above it - inefficient queries, too many transactions, inadequate indexing or whatever.

Unless I had a genuine use case Postgres could not handle, or some kind of disaster / recovery mode it couldn't do, or certification it didn't have then I don't see any reason to choose anything else. Basically default to Postgres unless there are reasons not to. The costs of Oracle and the burden of checking compliance are so horrific that I just don't get why anyone would choose it unless they absolutely needed it.

Now you can have all the fun of maintaining 2 sets of credentials and 2 consoles for every single environment you run. The only upside is potentially you can play the two clouds off against each other if one of them offers a better deal on some service than the equivalent in the other.

Oracle the database wouldn't disappear even if the company did. Somebody would buy it and but if customers thought Oracle squeezed them hard, then the next owner is going to squeeze harder. Just like Broadcom did when it bought VMWare.

Postgres is very good. Is it "enterprise"? Probably not in the way Oracle is but it can scale vertically and horizontally with replications and shards (citus extension). It's an excellent database that is pleasant and unsurprising to work with and has a lot of useful features. I expect that a lot of customers who paid through the nose to use Oracle could have used Postgres if they chose to - not necessarily for the super massive deployments but mid tier things. Migrating away from it might be disruptive but if there is demand for it, then tools and companies will pop up to assist with the process.

Even if the malignant company that birthed this dollar sucking monstrosity dies, the eponymous database and things like Java & Solaris will be bought by someone else as a going concern. And probably with an even more evil licensing model. Witness what happened to VMWare when Broadcom bought it.

Not that I particularly care but I'm sure it will bite a lot of companies who'll suddenly wonder why they even needed a commercial database for the workload they used it for. At least there are open source versions of Java, but anyone caught on Oracle if the company collapsed or was sold off is royally fucked.

Yeah but radio ads are generic and not 2 way conversations. They are not an asking what you are saying and twisting it towards explaining why you need product X.

It will be as reliable as asking a used car salesman for advice. Somehow it's gonna be advice about how a car would for me

All these places are streaming terrestrial TV & sports from god knows where but almost certainly not legally. Things like soaps, rugby, football, hurling etc. Bars are packed out when big matches are on. If Greece starts cracking down on this stuff these places are screwed.

An autonomous drone has to be able to cope with cables, pylons, trees, plastic bags & other blowing detritus, flag poles & flags, wind turbines, terrain, buildings, birds, light aircraft, drones, wind, gusts, down/up drafts, frost, heat, cold, strong sunlight, reflections, snow, rain, hail, lightning, etc. It has to be able to do this day or night. Drones also have their own failure modes on top of all this.

If a drone can't avoid obstacles and fail to safe then it has no business whatsoever flying near populated errors. Not to mention that even if did operate safely then it's still a noise nuisance. It would be very sad indeed if people start taking potshots at these things or griefing them.

I'm sure the reason boils down to some executive talking BS about "synergy" - that if they buy Arduino they can force boards with their microcontrollers down the gullets of grateful enthusiasts and educators. More likely to happen is enthusiasts & educators decamp over to Raspberry Pi land assuming they haven't already.