arrowleaf

On my 2023 Subaru the infotainment system routinely crashes on startup, I think if I plug my phone in too soon after turning it on. I'm sure the engineers at GM (and those at Apple working on Car Play gen 2) have worked through all of these issues that have been plaguing infotainment systems since they were introduced. /s

With most vehicles being 'fly by wire' nowadays, do regulators like the NHTSA and FMCSA have their own software engineers / technology experts reviewing what's going into these vehicles?

When 346 people died on the 737 Max 8s there was headline news speculating on the role of new software. The FAA had a lengthy investigation of Boeing and we saw pieces put out by experts like IEEE opining about what went wrong and how software played a role. Boeing paid $2.5bn to avoid being put on trial.

45,000 people in the USA die in motor vehicle incidents every year. How many hundreds if not thousands of those were caused by software faults? Is the software in sensor-based safety packages working as flawlessly as flight control systems?

I bet more than 346 pedestrians have been killed by drivers of Subarus that have EyeSight packages, drivers who think EyeSight might make a difference in them not hitting someone if they aren't paying full attention to the road.

Seriously. I can't believe that all of these auto makers got bailed out just so that they can screw around with stuff that will never be their area of expertise. Where is the innovation on gas mileage, impact safety beyond making the vehicles larger, getting decent diesels that can meet EPA regulations for passenger vehicles, etc?

Instead automakers keep on reusing platforms that are hardly different from their previous generation and tack on software-based 'innovations' like they're a tech company. Usually for the worse like in this instance, replacing the (hardly) tried and true interfaces of CarPlay and Android Auto with their own software. Before this it was switching out climate control knobs for on-screen which don't work with gloves, throttle profiles which will never be used, etc...

Yes, if you have downstream services that need to make use of your data, you would provide an API. That way you can manage the number of connections, how the queries are formed, any sort of encoding/decoding, authorization, caching, etc. Definitely not limited to read-only.

Have a single application clearly owning the data. All access to the data is done through that application. Why would you _want_ multiple applications deciding how to handle data in a common store? What about access authorization to sensitive data?

Implementations with skin in the game are still innovating and improving the 'tried and true' relational database model, IMO. PostgreSQL, MS SQL, MySQL, etc. consistently churn out new features that actually make sense, like FK constraint support for sharded/partitioned tables relevant to this discussion.

And then, don't just drop the constraints because you read an article about how FK constraints don't work with a distributed relational model, take a careful look at the docs of whatever DB engine you're using. PostgreSQL 12 for example supports foreign keys with either side being a partitioned table.

The only time I could see it really impacting performance is during bulk operations. But that's where making FK constraints deferred and wrapping the bulk actions in a transaction makes more sense.

My assumption (with modern applications!) is that nothing but the role directly owning the data will access the data. The development and DBA teams will likely have a role they can assume after performing a carefully-audited breakglass procedure to use in an emergency (rare) or to fulfill audit tasks. At least in my org this is a well-known problem with legacy applications sharing databases. Limit access to the database to a single role, used by a single application, and you absolve so many issues.

By soft deletion, do you mean you would want to ever re-instate the record and make it valid again, PK/FK constraints and all? Or tracking deleted records? For the latter I use audit tables + triggers to track the changing values.

At a basic level you could duplicate the entire record into the audit table on every action, e.g. the audit table would look like `audit_id | record_id | user/process_id | action (insert, update, delete) | timestamp | ...<record rows>`.

You can optimize it to not duplicating column values unless necessary. On inserts you only need the metadata of the action. On updates, the old value of columns with changes goes into the audit table. On deletes the whole record goes into the audit table.

There's also lots of track included that hasn't seen passenger trains in 30-50 years. The only passenger rail in Idaho is in the panhandle. I'm not sure if Helena has any passenger traffic either.

All of their messaging, Ilya's especially, has always been that the forefront of AI development needs to be done by a company in order to benefit humanity. He's been very vocal about how important the gap between open source and OpenAI's abilities is, so that OpenAI can continue to align the AI with 'love for humanity'.

I need to take another look at this. I've recently started working with RDF data on a scale of ~50 million URIs not including properties and statements, there's a ton of suplicates in there. I loaded a 10k entity subset of this and OpenRefine found all of the duplicates I had manually found plus others that I guess were 'similar'. Really cool, but it crashed when attempting to merge entities together. I've got a pipeline transforming the original JSON dataset to RDF, maybe it would work better working with the looser structure. What scale of CSV data did you have?

Serious question, is Red Hat a risk to the Quarkus project? I haven't started any new applications on anything but Quarkus for a couple of years. I would love to use Helidon, but the features and production stability just aren't there yet.

I love VS Code, but I will never be able to work with Java code outside of IntelliJ. Hopefully this extension is a bit more stable than the Red Hat one, whenever I've attempted to use VS Code for Java it's hit or miss whether plugins or the language server crash.