TCP uses a mechanism called slow start. This starts out transmitting information at a slow rate, and gradually speeds up until it reaches the maximum the network can handle. For long-lived connections this is a really good approach, but for small requests such as a web page, it will probably transmit the entire file in the slow phase. If you then kill that TCP connection and start another for the next file, then the next file goes through the same slow start process.

HTTP 1.1 will re-use a single TCP connection to receive multiple files one after another - but that is still slow since the server has to finish sending one file before getting told what the next file is, and then going and loading that file from storage into memory and prepping it for sending. It would be quicker if you could say 'send me the following 20 files' at the start, and then the server could keep busy by loading the next file while the first one is being sent.

Because for HTTP you have multiple streams going to the same host, so you can share the TCP set-up costs across all of the streams. Starting multiple TCP connections to the same host would be slower.

By implementing multiple streams over a single TCP connection you only pay the TCP set-up cost (slow start, MTU discovery, etc) once. If you break it out into multiple TCP connections then you incur that cost for each stream. Since all the streams are going to the same host, you might as well share the same TCP connection.

Because drones at the moment are much closer to the bomber role than the fighter role. They are armed with air to ground ordinance, they are turbo-prop driven, and operate only in uncontested airspace.

The current generation of drones would be mostly ineffective in a battle against an enemy with an air force of their own. Even a 3rd generation fighter aircraft could take out modern drones without trying very hard, and I believe they can't presently arm drones such as the MQ-9 Reaper with air to air missiles (although they are working on that).

a large hole next to a large hill that you dug out yourself? I wonder if there are any studies on the practicality of man-made hydro storage,

Put the turbines half way down. When you need extra power, run the water through them into the lower half of the the hole. When you have excess wind / solar power, pump the water back up to the top of the hole.

Actually math is probably the better option than pictures. I II III IIII IIIII I+II=III II.II=IIII III-I=II

Have you ever written out a long and detailed response to an on-line post (or an email in a web based email client), and submitted it only to find out the network link is down, or the remote server is down, or you have moved out of wi-fi range, or any of a thousand other reasons why you can no longer submit the form?.

With local storage, the client side javascript can handle this for you by saving a draft copy locally, and then later when you have your link back again you can send it successfully.

What if the fans were mounted in the middle of the wheels, and tilted from vertical to horizontal when needed. There would need to be some kind of stand on the bike that held it in place while the wheels tilted around, and some handling of the rear wheel chain/shaft linkage, but that shouldn't be hard to do.

I generally believe this is a major factor in underestimation, Even a "good" manager will unconsciously apply pressure to produce optimistic estimates. I was once asked how "accurate" my estimates were, and I said +/- 15%. I was told to go away and work out a "3%" estimate. I added 12% and gave the numbers back - they went nuts. They expected the same numbers but with a promise that they were more accurate.

Unfortunately not, because you never got any work, so you went out of business. The company that lied about it has managed to suck another 50% out of the client - who has never heard of the sunk cost fallacy, and then used a bit of money from another contract to deliver something that has 75% for the functionality. The client is mostly happy because they asked for the world and got something that was almost good enough.

Actually Google is deducting the cost from their tax bill. The catering company is deducting the cost of the food, staff wages, and all other business expenses from their income. The only parts that probably gets taxed are the income of the catering company staff, and any profit that the catering company makes. Since catering typically has very low profit margins and pays minimum wages I expect this tax to be rather low.

Yes, but your mom doesn't get to claim the cost of buying and preparing that hot pocket off her taxes. Google does. So not equivalent.

A restful back end driven by an html/js/ajax client is my favorite architecture as well, however if you want to do a server-generated dhtml style application, I would suggest you have a look at Stripes - It's everything that Struts should be (and isn't).

The only real issue in this whole debacle is the large loss in battery charge while the car was parked overnight. Looking at the graph that Musk posted here I can see the battery charge taking a steep dip right as the car is supposedly parked. The graph of remaining miles shows it even more clearly - obviously the computer was extrapolating from the sudden battery charge drop.

So, what caused the sudden drop? The speed graph isn't fine enough to determine of the car was driven, and although there is a cabin temperature spike, the reporter says that happened the next morning when he was told to run the heater for a while. The engineers were obviously thinking it was temperature related, and thought that with a bit of "conditioning" it would all be ok. Thus the suggestion to run the heater, and to slow-charge. Finally the assumed the computer had it wrong and told him it was ok to drive - and were probably wrong.

So, the only real conclusion left is that the battery actually lost charge overnight. Did Broder sabotage the result by running the heater longer than claimed, or drove around in circles (again) to run it down, or maybe he just left the headlights on overnight?. We'll probably never know.

The alternative is that the Tesla batteries discharge substantially when not being used in cold weather. That should actually be pretty easy for someone else to test.