Hacker Newsnew | past | comments | ask | show | jobs | submit | tylerflick's commentslogin

I have one of these. I would recommend looking at something else.

I hate managing clients certs, but I doubt there’s a better way.

You can (within reason) calibrate these cheap CO2 easily with esphome and an outdoor air reference point overtime. I’ve had to do it for thermometers as well.

ROCm is a mess. I gave up on it and decided to run OpenCL on Vulkan: https://github.com/kpet/clvk

Thanks... I haven't really followed ANY AI stuff up to this point, other than awareness that it exists... so have to say, my shallow dive yesterday was a bit off-putting to say the least.

I'll dig into this over the weekend when I invariably try again.


> Microsoft pile drove them from early dominance to niche by undercutting them handily

And hiring Anders Hejlsberg


Microsoft did a lot of bad things over the years, but Borland drove themselves over the cliff on their own. Instead of focusing on developer tools, they wanted to reinvent (and rename) Borland every few years in the 90s.

Bad management, bad decisions, bad products (Delphi 7 was peak). MS had nothing to do with that. And I'm sure Anders made a right move to abandon the sinking ship.

I'm still pissed at Borland for all those bad moves.


It’s very difficult to make money in developer tools. Microsoft could easily squeeze Borland by simply making MSDN tools free. Borland tried to diversify with databases, word processors, and spreadsheets, but Microsoft countered with Office, trying them altogether, and it became the default in every single business. Borland had great technology and was super innovative and I used Turbo C++ and TASM for years. But in the end, they just couldn’t find a cash cow market to keep them afloat.


> It’s very difficult to make money in developer tools.

Just to be clear: we are talking the 90s here. Everybody was charging for developer tools (). MSDN was not free, far from it. From today's viewpoint where every compiler imaginable is free and the tools are better than ever (except there is nothing like Delphi and VCL), the 90s were a heaven for tool makers.

I'm talking about the Windows ecosystem.


Correct, but Borland didn’t die in the 1990s. That was its heyday. As I said, I used Turbo C++ during that period and I spent good money on it. But the tools commoditized and Microsoft eventually made MSDN basically free in the 2000s (there might have been some nominal charge, but it was low). And that was when Borland eventually got acquired, in 2009.


The could have been the Jetbrains (before Jetbrains existed) and even bigger than Jetbrains.


I remember thinking back then that Jetbrains had their pricing right.

Not free but low enough so that invidual developers and companies wouldn't think twice about bying a license.

Borland/Inprise/Codegear/Embarcadero just priced themselves out of the market.


Just offload themselves to JetBrains, and have Delphi opened.


Open Sourcing Delphi along with Commercial licensing to Continue development is on my wish list if I ever become a billionaire.


And yes, naming themselves inprise was peak 90's wallstreet cringe.


Borland decided that they should target management instead of the developers as their focal point of product development. They ignored the Web for Delphi and decided that web development front would be covered by JBuilder, a paid and slow evolving product that could not compete against the fast iterating and free Eclipse.


Thanks, got a good laugh out of this one.


The photo in the article looks to predate the mass adoption of cars. Maybe it was because of new engineering capability + attempting to avoid the elements?


It's probably both. The picture in the article is very early days and was then a proto-PATH that was probably first built to deal with an acute problem (probably crowding on the sidewalks as the streets started to be taken over by vehicular traffic).

Toronto's PATH doesn't have a central control or planning system. It is literally a series of 1-1 agreements between buildings that build tunnels under streets to connect themselves. The main benefit to each building is that they can charge retail rents in their basements for through traffic. The system map was terrible and even had a planned route prematurely showing a way to the Eaton centre from the south for a building that was left uncompleted for almost 30 years (work stopped in the early 1990s recession and was only finished right before COVID hit).

Avoiding the weather for commuters coming in on the subway and GO train (suburban commuter rail) was a nice benefit, though only very recently was Union station fully separated from the elements. The one problem with PATH is that the shops are completely targeted to 9-5 work commuters, particularly to finance workers; think coffee/business suits/lunch/etc. Though portrayed as a giant mall, almost all the shops are closed on weekends and don't stay open much into the evenings (Montreal's is more dynamic comparatively). COVID of course upended the business climate of the shops, too. Some newer condo towers have been connected and there is some very early signs of something more dynamic, but the towers seem to still be holding on to the idea they can charge pre-COVID rents. My personal opinion is they should be seeing this as a loss leader to convince people to want to come in 5 days a week (cheap, good lunches, etc).

Anyways, we'll see how it continues to evolve.


The first one was built to connect two of the same company's department stores (Eaton's, now defuct, but once an iconic Canadian chain).


This is not the first historical inaccuracy I’ve noticed from this site.


> I've often wondered how (in)expensive you could make something with a better built-in screen, the ability to play VP9/AV1 video and opus samples and Bluetooth to connect to mouse / keyboard and PC as a net and storage device.

I’ve been doing this, and the answer if you have a 3d printer is $10 US(pi zero 2 w) + the cost of whatever display you want to use. This won’t do AV1 but it’s good enough for my use cases. I’m slightly cheating on the cost as I have plethora of LiPo batteries in reserve.


I’ve been wanting to do something like this to make a modern iPod / audiobook reader.


For an iPod style device you could even go lower down the line and use Pico W along with the addition of some i2s/amp combo.


If you’re in KC be sure to check out the TWA rocket in downtown too.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: