When ideology trumps pragmatic design

Real differentiation, adding real value to something, is often hard to do. Fundamental changes often take time, and are often incremental in scope, so they don’t break everything.
That is, unless you are so completely convinced that your way is better, that you try to force the market in that direction. Sometimes these gambits work. Sometimes they don’t.
This is about one that did not work.
I am convinced my Mac OSX laptop may be the best laptop I’ve used. Its like a Linux box with fewer screw ups on random things. I have gnome terminal, I have all my aliases, I can do everything I need to do from it.
I have a mac mini on my desk. It doesn’t get used as much as I want to use it. Because my linux desktop is, fundamentally, far superior in all aspects. Including usability. Yeah, that is what I said. Linux on a 16 core 48GB RAM system with 6TB of very fast disk is freaking awesome. The mac mini is … nice … but it can’t hold a candle to the other box. That is the best desktop I’ve ever built.
In that desktop, if I want to expand it, I can add PCIe cards, and they will largely just work, and work correctly. No silliness, not messing with configs, it will just work.
On the Mac’s, if I want to add a PCIe card, I have to add a card cage. Of course, I could use the old MacPro box, and it has room for older PCIe gen 2 cards. But nothing for PCI gen 3. And the new cylindrical Mac pro has no room for any PCIe cards.
To expand it (or the laptop, or the mini), you need an external card cage.
Which, when you talk to the few remaining PCIe card vendors that support Mac, and you ask them about their plans for Mac pro (the cylindrical bit), most shrug their shoulders. What can they do? There’s not real Thunderbolt to 10GbE, or 40GbE, or FC converter. There are Thunderbolt to PCIe card cage, and you are running in a daisy-chained link environment. Which means high bandwidth/capacity cards, like, say, 40GbE, advanced FC … etc, can completely flood the wire.
Thunderbolt is basically a serial PCI bridging connection, so in theory you could get better performance from it by “bonding” these links together into multiple parallel connections. But the card cages don’t quite let you do that.
Yeah, the thunderbolt bit is cool. Nice bit of tech, and its faster than USB3.
But its about as useful as USB3 for connecting high bandwidth to other units. That is, it isn’t that useful.
Someone clearly did not think this through. There are no thunderbolt switches that I am aware of in market. I’ve been looking and have hoped to be wrong, but I don’t see any. There aren’t any bonded thunderbolt to 10GbE or 40GbE. We hacked the latter up, but I am not sure I’d recommend a customer deploy it.
Way back when I started thinking of building a company like Scalable Informatics, I read a short story by Arthur C. Clarke (yes, him of 2001 fame, amongst others) named “Superiority“. You can read the whole story at that link. I was amused initially by the forward which noted

Soon after publication ‘Superiority’ was inserted into the Engineering curriculum of MIT – to warn the graduates that the Better is often the enemy of the Good – and the Best can be the enemy of both, as it is always too late. — Arthur C. Clarke ([ACC, p. 395]).

That amusement turned to understanding, as I watched how people with inferior old tech won against people with superior new tech, time and time and time again. It didn’t always happen, but it happened enough to be something to be studied.
Proper execution can obviate the advantages of the new and different. This is something that may have been lost on the Apple folks when they started working on thunderbolt. I can tell you it is poorly understood in many areas on SV, but thats a discussion for another day.
There are always consequences to actions, and costs both real and opportunity, to decisions. In this case, without a PCIe card capability, Mac’s won’t be able to play well/easily in our high bandwidth massive capacity future. Maybe this is by design. I don’t know. But Thursday, I go out to a customer site to install a Unison machine talking to two Linux boxen and one Mac. That might not be able to keep up with the Linux boxen. So I am looking at alternative mechanisms for the customer, including moving off of the Mac.
This calculation, this migration potential off of the platform isn’t because its a bad platform. Its a great platform. It just can’t do the things its competition is doing very well. Its competition is superb at execution. Apple is superb at bringing new and cool things to market. But new and cool that doesn’t solve the need for these customers, in a market that Apple all but completely owned …
… look at the object lesson of SGI to understand the impermanence of market leadership in a space. Hubris, crappy execution, massive internal wars, … all at a time that this emerging platform from this Cupertino company was starting to come into vogue.
I am not claiming Apple is not long for this world. They make great kit. But at the same time, like all other companies, sometimes they mess up. And do so in profound ways. Ways that completely open the door to competition, and effectively exclude them from playing there. I hope it was unintentional. They need to rethink the no-PCIe card design. Cool doesn’t matter for squat if you can’t do your work on it.