I’m not blind to the fact that Windows has a terrible search experience, but I won’t say it’s the worst. Out of the box it is fucking useless, but if you actually take the time to index specific paths and make use of the metadata options in Explorer you can actually get some decent results in reasonable time.
Apple had the right idea with Spotlight, it’s just sad that both parties can’t properly integrate the functionality into their file managers.
My search needs are mostly covered by Voidtools’ Everything, which is able to scan through the whole NTFS partition in a matter of milliseconds for realtime results. The caveat being that you have to know the name of the file you’re looking for. Otherwise I just use Powertoys Run for search and application launching needs, which is what Spotlight could have been if it was made by passionate nerds.
I do realise this makes my argument lose it’s bite somewhat, but it comes down to user experience. MacOS has a terrible out of the box experience that can’t be fixed, use something else instead. Windows has a terrible out of the box experience that can be fixed to some degree if you take some time to learn how it works, but you can still opt to not bother with any of it and use something else instead.
Linux was always going to be the clear winner here.
Now for the dock icon strategy: try doing that repeatedly with multiple instances/windows from the same application and compare the experience to the “never combine, show labels” taskbar in Windows. I guarantee there is going to be a clear winner in terms of usability. As always, under Linux is not an issue because you can just do whatever the fuck you want.
The troubleshooting bullshit a pattern seen in all of Apple’s products. They have a habit of hiding all important information in case of an issue, and I have had this complaint about all the iDevices I’ve had the displeasure of touching over the years. iPhone update failed? Tough shit, have a red message saying “something went wrong, try again”. Fan controllers randomly stuck on 100% speed on an iMac without the ability to get any information about the sensors. None of the system tools provide any information beyond the bare minimum. I’ve come to a point where I just refuse to help any family member having trouble with an Apple device because it turns into a multi-hour wild goose chase.
I’d argue having an overall stable experience with the occasional vague issues that can’t be troubleshooted is worse than having slightly more frequent issues that actually lead to a solution. Apple’s products by design teach the user nothing about technology because there’s no entry-point to the knowledge itself. Meanwhile Windows, while flawed, does provide the user with an incentive to learn about proper maintenance and troubleshooting which leads to more competent users overall.
Last but not least, the command line: I love my experience using package managers on both Linux and lately also on Windows with winget. It has quickly become the main way I install and manage most utilities. MacOS has options but none of them are integrated neatly and have to be installed separately.
Even on Windows I use command line utilities where I can and GUI functionality where it makes sense. While the realm of possibilities is not as broad as the GNU/Linux world provides, I at least feel I have a great deal of control over what I do. MacOS is an impostor that has murdered a UNIX distribution and is wearing its skin. The terminal experience feels like it’s a remnant from the early days that they never bothered to put any more love into but they can’t get rid of it either. Just like some of those Windows 3.0 components you can still find in the modern versions of the OS.
I’m 100% done with Apple and their products because they make everything I’m trying to do slightly more difficult and annoying than the alternative. And those are just limited to my issues with the way they do software. I also have very valid issues with the way they design hardware and with the way they conduct business (ethics, monopolies, their overall effect on tech markets in general).
Don’t be discouraged, it doesn’t come naturally and there is good reason to do so. The Scots are generally awesome people and the world needs more fer’s, aye’s and nae’s in general.
Jus’ expose yerself tae sum more Sco’ish and ye’ll be jus’ fine, lad.
From personal experience working in a Microsoft ecosystem, it’s mostly a matter of being able to hire the right people.
There is a near-infinite source of IT workers that have some expertise with Microsoft software and services. And those kinds of numbers simply don’t exist for the Linux world, especially with all the different configurations out there.
Medium-sized organizations have to employ a strategy of throwing enough idiots at a problem in order to keep things running. This also creates some of the issues they need the idiots for because no one has detailed knowledge of how things work.
My attempts at proposing a linux-based application server have been met with all sorts of “but our domain policy”, “we can’t guarantee continuity”, “none of my people know how to admin this stuff” type responses.
It definitely is a matter of mindset, but there is also a big commitment to make if switching systems to Linux. And that is a choice managers will only make if the benefits are clearly illustrated in a businesscase.
I haven’t dabbled that much in PCB design but I have seen some good things in KiCAD. All my electro engineer homies assure me Altium’s the way to go for now though. Most of them also happen to be big F(L)OSS nerds so I’m curious to see where KiCAD goes in the future.
FreeCAD is an awesome attempt at building a parametric CAD modeler, though it will need a lot of polish to be usable. Especially on the UX side of things the software could do with a lot of improvement. As far as I know the most difficult part to program for parametric modelers is the actual geometry kernel, which is why so many modelers are based on Parasolid, including the recent hybrid modeler Plasticity. For a F(L)OSS parametric CAD modeler to truly succeed some genius needs to build an open geometry kernel that performs at least close to on par with Parasolid. But that takes a special kind of autistic in order to achieve. Either that or the engineering world needs to collectively decide this needs to happen.
As much as I hope FreeCAD becomes the open source alternative everyone is looking for, it is trying to be everything at once and that might be too ambitious for the current state of the project. I’m secretly hoping we also get a new project sometime soon with a smaller scope.
Now comes the hard part of defining all the Eeveelutions.
I feel like there are a few very distinct regional accents, but I’m having trouble coming up with the right distinction from the top of my head.
There’s New England, the south in general, New York, Chicago which immediately trigger my brain to think of a very specific accent. Surely there is more to it though?
While I’m not exactly an expert user of AutoCAD (my background is architecture, industrial design and full stack development), I know enough about the software where I can tell it’s based on a lot of legacy spaghetti code.
It’s the same for Solidworks, which I know through and through, including the shitty VBA scripting environment. My CAD teachers always used to say the software is built like a wooden playhouse, which has been extended over the years to include a second story, a slide, a swingset and a roof extension. But underneath it all, it is still the same “don’t fix it if it aint broke” codebase that Dassault has taken their chances on since the '90s.
The second someone invests any kind of money into an open source alternative, the way Blender has done for the mesh modeling industries, both Autodesk and Dassault systemes stand to lose their respective monopolies on 2D and 3D CAD.
But the trend is not limited to CAD software only, it is also highly prevalent in software providers for governmental tasks. Most of which sell the same products for years without iteration on their codebase. The result is that government organisations have to deal with shitty software that requires their individual users to connect to the database (yes, you heard that right, every user has to manually input database credentials that include all grants on all of the relevant datasets). Most of these cronies are reselling badly thought out software, where they’ve outsourced the development to third-world shitholes. Is is a goddamn miracle that there aren’t more major incidents with government organisations.
The only solution for this kind of bullshit is open standards that encourage an open source approach to these kinds of critical applications. Where more parties are actually encouraged to build their own software and where the businessmodel is built around being a service provider and not a magical black box salesman.
If you’re able to stop worrying about generating revenue based on your intellectual property and focus on generating revenue from the service you provide, surrounding your product… you’ll automatically build a better product.