Intel Compute Stick review: Second time's the charm
engadget.com
engadget.com
Something I learned soon after I launched my first hardware product:
People still have an instinctive physical relationship with products that must be the result of evolution making a connection with heavy or solid = more-food/safer/better or something like that.
I remember devoting a ton of time to making my first product as light as possible through sophisticated mechanical design and attention to component choices in the electronics.
During my first trade show in Europe, this Italian guy came by multiple times to look at the product. He asked for a price on his second visit. It was USD $8,000. The next day he came back one last time and asked: Why is it $8,000 if it is so light?
Absolutely blew my mind. Was not ready for that question at all. Upon return to the US I directed our sheet metal vendor to switch from aluminum to steel of the same gauge. We also added a small steel bar riveted to the inside to increase the weight a further. In all, I think we easily added about five pounds to the thing. Now it was "worth" $8,000.
Over the next three years sales and comments during sales presentations proved, in no uncertain terms, making it heavier had been exactly the right move. People would make comments when they picked up this damn paper-weight such as "wow, feels solid", "wow, lots of tech in there", "wow, I know why it cost so much now".
It was amazing to see how many times the moment the deal was done coincided with the instant they picked it up.
Amazing.
Interesting -- what type of product was this?
I've designed equipment with small, lightweight boxes that sold for similar prices ($5,000 in my case) and didn't run into that phenomenon in my (very small) niche, but I've definitely seen it in action in contexts like EEVblog teardown videos. Coincidentally, I'm working on a new design right now, and one of the items on my to-do list for this weekend is to decide whether to go with aluminum or copper for some internal shielding. :)
Now you know why Beats cost so damn much.
I have to consiently suppress this feeling a lot of times. Also, there's a relation of bigger -> more powerful, but this one is easy to suppress on portable stuff.
Is that still the Kangaroo? I can only find one model listed on their site and I want one...
I spent an afternoon once getting Linux to run on the Lenovo stick and an HP Stream 7. I returned them because I don't ever want to have to do that again. I can't even picture using one of these on a daily basis.
Particularly, to be used to connect to DaaS.
I can totally understand how it serves Intel to squeeze the most out of corporate buyers, and also it allows one to get a bargain from time to time... But I think in general it's just confusing and annoying.
Not criticizing your post at all. Simply making the observation that what we might call a "dumb terminal" today is a million miles away from where this term originated. Just like the "film industry" is really the "digital image industry" today.
I remember running AutoCAD 1.2 on an 8086-based S-100 CP/M system with the addition of an 8087 match co-processor card, 640K of RAM, 1 MB on a RAMDISK card, a tablet with a puck that used a magnetic coil to sense position and, yes, a DEC VT-100 "dumb terminal". Sometimes I had to unload stuff from memory just to be able to plot. Fun times. Funny that today developers think stuff like vim is cool. We couldn't wait to get off those damn terminals and use "real" editors.
Who knows what Ctrl-K-X belongs to?
Back in those terminals' heyday, they were "smart" terminals! Dumb terminals could handle text output (sometimes to paper), but had few if any control codes and certainly not the extensive VT100 escape-code set for cursor control, formatting such as bold or underline, etc.
Who knows what Ctrl-K-X belongs to?
WordStar! And joe.
WordStar?
That also explains the way ACAD works. Had the program been created in the post-terminal age it is likely it would have been a point-and-click application, like, say, SolidWorks. Instead, they relied on commands such as "line", "circle", "pline", etc. typed into the console. You could also type the parameters rather than clicking around.
ACAD also came with a version of Lisp (AutoLISP) you could use to extend its capabilities. In general terms the approach was to use AutoLISP to create a DSL that made drafting much easier for your particular application.
For example, I had a set of programs that created a DSL I could use to layout circuit boards more efficiently than through raw command drafting. If I remember correctly, I could type "dip" and my code would ask a series of question to ultimately place a, say, DIP 16 package with certain size pads and holes on the current layer.
The degree of automation one could be achieved through AutoLISP was remarkable.
That pattern of cmd driven UI was very common in those days. I think just about every EDA package does that. However, instead of Lisp, they used TCL.
Thanks for sharing this. I think it is always great when people achieve powerful automation.
This post has an example on how to limit a process (in this instance it's Matlab) regarding it's memory use an the CPUs it will allow to be scheduled.
To limit the CPU further, you might want to use the "userspace" cpufreq governor and allow down your CPU clock.
This PC is a little weaksauce to be running Windows, but for typical Linux applications it's overkill. I don't know what its niche is, except as a terminal server or digital sign -- not exactly growth markets for home users.
Businesses will eat it up.