TTFB role in the overall time to a usable web page has dramatically decreased. Instead of being the primary driver in display of mostly static sites, the role is smaller now due to the increased compilation/execution time of client code.
What tld? It’s likely premium but they don’t get that info back until they call into the registry. Comparing prices between registrars is good practice as well.
Not specifically Ventura, but anyone having issues having os remember monitor position? My Mac mini running Monterey can’t flips the monitor on login/unlock more often than not.
I had never had issues with external monitors until Ventura on the M1 Air.
Now, each time I plug it in to the monitor whatever will happen is pretty random. Sometimes it gets the full EDID from the monitor and everything is as I want it to be, including resolutions, refresh rate and positions. Sometimes, I get partial EDID and only some refresh rates available (my monitor goes up to 144Hz, sometimes macOS can only do 60 or 75Hz). When that happens, resolutions and screen positions may or may not be ok. Most of the time the laptop screen is set to mirror the monitor which is pretty much never what I wanted.
Other times it just doesn't properly detect the monitor which stays blank until I disconnect and reconnect - at which point any of the above paragraph might happen.
More rarely, the computer freezes and I need to force shutdown.
This isn't the topic of the article so I haven't included it, but I have taken several "solid steps" to optimize this application throughout its year and a half of being in production.
This article is about one of them only, how I designed a test system to evaluate the two databases head to head.
Query planning seems like it should be the topic of the article. SQLite and PostgreSQL do their query planning very differently - not to mention that SQLAlchemy is probably generating different queries. I can't really fathom how one can meaningfully "evaluate the two databases head to head" while seeming to actively avoid how they plan/optimize/execute queries.
I don't have access to the code used in the article. Therefore, no amount of research on my part is capable of answering why PostgreSQL was faster than SQLite in the author's situation - certainly not without the queries and the plans / EXPLAINs thereof, those being the specific thing I and multiple others have suggested as something the article could've provided and which would've made it an actually interesting head-to-head.
Meanwhile, if all you have to offer in response to such a suggestion is "hurr durr do it yourself", then why bother commenting at all?
This is unlikely. When a registrar goes bust, the registry can take over, assign a replacement registrar by buyout or alternate agreement. See “Registerfly” case as an example.
In some cases though, the registrar only pays the registry for 1 year at a time, making the money their customers pre-pay effectively a low-interest loan for the registrar. If that registrar goes bust, you'll be able to transfer your domain to a different registrar, but you'll lose the money you paid (i.e. loaned) to the original registrar.
There's also another risk: if you decide after 2 years you're not happy with your registrar, tough luck, you've already paid for 10 years.
That's not usually a problem. You can transfer a registration with N years remaining, the new company will charge a one year registration and you now have N+1 years covered.
The only time it becomes a problem is if your new registrar doesn't support N+1 year registration - your 8-year example means you need to find someplace that supports 9- or 10-year regs.
Read the first part of my comment again. Some (many?) registrars don't actually register the domain for N years at once with the registry, even if you pay them for N years.
If they don't pay the registry, wouldn't it not show up in whois? If it doesn't show up with the expected expiration date in whois shortly after registration/renewal, I'd be demanding it get fixed.
Also seems that messages land in the inbox first, and then are sometimes filtered, making the gmail web ui jump around as whatever filtering mechanism catches up.
