Re: mmm
CF has a pretty good reputation as an engineering first culture. TBH, the only "error" was not notifying the customer.
Guessing the escalation process being added will address that.
15 publicly visible posts • joined 23 Mar 2017
I'd definitely bring computer, db and search in house. And would replace S3 with Cloudflare R2, for the usage that's public facing.
Search services seems to be the most singularly overpriced SaaS out there compared to self hosting elastic or similar.
I would probably replace the Dynamo with Cockroach which will increase usage options and slightly easier to manage than Cassandra.
Depending on the total usage a couple racks at 2 hosted facilities should be a cost savings and likely better performing.
It's funny, VPS makes the most sense at small scale... Cloud at mid so you can save on personnel, but hitting the point where self host in a rented days center eventually takes hold.
When I was younger there are points I did have two jobs. But it wasn't a secret and was very long hours.
If you can fill expectations for two jobs at once, more power to you. I know there are jobs that are simply under provisioned out there. I've considered it. I could take two lower-mid level positions and make slightly more than the very senior positions I've been in the past few years. So who knows.
If your job performance sucks, then your performance sucks and the reason shouldn't matter.
Not sure that I'd push for any new projects using Microsoft SQL Server over PostgreSQL, But having sequel available on Linux is really helpful especially in containerized scenarios (docker) for testing and local development.
Especially considering a lot of windows service/web projects developed in the last 5 to 10 years can easily be migrated, or just recompiled to a Linux host environment.
TIP: When migrating, convert all dates to transmit and store as UTC unless paired with a location. Convert to and from local at the client. Configure servers to operate as UTC/GMT
And this is precisely why nearly everything I've developed in the past decade can run in Linux. Even apps hosted on azure. I abhor lockin, some things are easier to swap out than others.
Other than AD for a lan, there's very little reason to run windows servers. For office/email, may as well rent it. I like some MS software and service.. But windows has much less of a place in my life.
The goal is 100x their current tech which from other comments seems to be 2+ decade old production of even older tech.
A combination of redundancy better fault Torrance and shielding could give a combination of current arm CPU design with some custom processors for certain ai tasking more than a chance. Given the state of Arc, it may be a good fit for the task... At least as a starting point.
The only place I would use SharePoint is transparently behind the MS Teams application. I like Teams, the user of does and wiki should generally stick to only the main channel for a team though. It's been very useful.
I worked with the first few versions of SharePoint and hate it to this day.
While I recognize that there are useful places for these features, they are most often used by "your computer is infected" ad scams. For that matter, I'd like to see JS disabled altogether for content more than two iframes deep. Look at ad network content and browser overhead sometime.
Even in a monster JS app, loading way too much JS, a single ad frame can dwarf it easily. I reviewed a web app for a friend that was behaving badly... It wasn't the app itself, it was the third party "chat with us" module in the corner.
It iris me to no end far more when integration plugins like that,. Or ads aren't coded with extreme care for the impact.
While I don't use a screen reader, I do get it. I spent years developing contents for e Learning with strict accessibility guidelines. It's not even that hard to just stick to sensible semantic markup with css and add a couple aria attributes here and there.
You're off course not surprised how ignorant the output from some devs can be.
And for those wondering, yes, you can make all this work in JS driven apps. Of course having to meet WCAG 2.0 for an app meant to visually process scanned documents, that's nuts (have had to do that).
There are ARIA attributes and other assignments that work perfectly fine with JavaScript applications. So yes, they do work. Some frameworks (mui for react, for example) take a lot of care to have the appropriate markup for controls, of needed you can still do partial or full rendering server side via next.js.
That doesn't even get into semantic elements or the proper use of css, which work perfectly fine with browsers and don't require J's at all.
There was a caching method in place supported in windows, but not in osx... Chrome on OSX didn't have the sluggish behavior on the fallback that it has on Linux, same fallback. Safari on osx didn't support the feature at all, so they turned it off for non-windows. It's actually a regression issue with Chrome regarding the fallback. That said, apparently it's corrected/on for the next release/update/patch and should be online soon.
Beyond this, not all functionality can be detected, especially specific browser/version/platform bugs via feature detection. The support agent could have been a bit more courteous, and should have forwarded the issue to the dev team anyway... apparently some devs picked up on the related story via Hacker News earlier yesterday. And they'll be adding linux more testing to the product in question (likely their automated regression suite).
In the end, I believe it wasn't malicious... I've seen MS devs go out of their way to help projects that only even connect to their stuff.