Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror

Comment A lot of issues (Score 1) 165

Disclaimer: I'm a so-called LEGO Ambassador, i.e. I represent my LUG (Lego User Group) to LEGO, but I'm not a representative and/or work for LEGO itself.

LEGO is very interested in this 3D printing topic and had a workgroup on this on the Ambassador forum. I did not participate in this workgroup, but I can give some of the results. None of them come as a surprise, if one thinks this topic over, though, so I'm not telling any secrets.

- for standard bricks, it is too expensive, and except for a few classical basic bricks, there are patent and copyright issues.
- for bricks that do not exist from LEGO, this may work, but color, clutch power, surface structure, and durability are nearly impossible to match with current technologies
- best use for 3D printed stuff is to technically link LEGO parts to other things, e.g. a RasPi case that can be connected to a LEGO technik frame, where color and surface structures don't matter at all, and clutch power does not matter that much
- A lot of 3D stuff is accessories for Minifigs, like tools, weapons, hair pieces, etc.

Basically, while there are thousands of 3D data sets for LEGO parts available on the net, actually printing a box of bricks to build is far from being practical.

LEGO uses 3D printing in their design process, but only for prototyping. They are more likely to cut and glue existing parts for the prototyping, though, as this is still faster and better.

Comment USB Serial - Too much brain for its own good (Score 1) 299

We have an application where we still need a real serial connection. USB serial adapters have too much brain for their own good, and don't cut it in realtime scenarios.

And from the embedded point of view, a UART can bed one with a handful of registers, maybe an interrupt, and a few lines of code, even in assembly. For talking USB, I need a whole protocol stack with hundreds of things I never ever need.

Comment Louis, just queue in! (Score 1) 378

What can be more palpably absurd than the prospect held out of locomotives travelling twice as fast as stagecoaches? - The Quarterly Review, March 1825.

Heavier-than-air flying machines are impossible. - Lord Kelvin, President, Royal Society, 1895

640K ought to be enough for anybody. - Bill Gates (1955-), in 1981

I admit that humans are going to Mars to settle. But there, human space travel will end. - Louis Friedman (1941-), Engineer, in 2015

Comment Back in school... (Score 1) 220

About 30 years ago, we already had pocket calculators in school. My elder sister, though, still had a book about calculating with a slide rule. I managed to get the book from her and a slide rule from someone else, and started exploring. Being the math geek in class, I could afford a bit of risk, so I brought the slide rule to a test where all my classmates used calculators (I still had mine in the bag in case the teacher would forbid using the slide rule). You should have seen the faces when I used that "plastic thingie" - and still was the first to finish, and later got the top score in class (again). Some even accused me of cheating, but the teacher was quite amused about it.

It was a fun experience, and I could even teach other people about it in university - we had a giant, working slide rule (a few meters long, probably an old demonstration device from a lecture room once) as a decoration on a wall.

Comment Young Talent - Lack of experience (Score 3, Interesting) 229

I'm one ofthe old dogs. I have to admit that App development and such is not my kind of things. But I do have experience. LOADS of it. When I see the fundamental mistakes by young "talented" programmers, it makes me cringe.

Just a few days ago, there was a kickoff meeting for a new project. This project needs multi-user support on the long run - everyone in the team admits that. And access control, with all its implications like "how to I check a password", "how do I store a password", "which kind of permission do I need to call this function". Which they never ever did before. None of them had ever heard of books like "Applied Cryprography". There is a copy here, on my shelf. Actually, it is my second book, the first was worn down due to heavy use. All they cared for was "Licence Management", but I'm not sure if they understand how this works properly. I offered them to ride piggyback on the existing licence management scheme I've implemented in my part of the system, but this was probably too unsexy, because it cannot add licences on the fly over the web, at least not "just so".

My experience tells me (and anyone who has been around for long enough) that any software that will need this kind of multiuser support needs to have this built-in from the very beginning. The very concepts of the software must be aware of the possibility that e.g. a call might fail for lack of permissions. Communication protocols must be designed in a way that they guarantee to a sufficient degree that one side has proper identification presented to the other side to be permitted to do this, and don't that. This is nothing that can be added lateron without SERIOUS headaches, problems, and, worst of all, risks. Windows9x was the living prrof of such a mistake.

Reply from the "young talent": Implementing multi-user is too time consuming at the moment, we will add it later. *FACEDESK*

Comment I am not exactly surprised (Score 1) 213

Math, and its application, never was a strength of the economists. Nor was or is logical reasoning or application of scientific methods.

Ages ago, I made a fake economics whitepaper. It looked like the stuff I've seen in their libraries, the text was equally braindead, and it contained a lot of made-up formulas. The formulas were completely irrelevant (as was the text), but if anyone had followed the pattern to read the "input data" in the text before and actually pass it through the shown calculations, he would have noticed a) that the actual result did not match whatever was written in the text, and b) that all formulas resulted in an 8-digit number (with the decimal point in varying places) like 1991.0401 or 199104.01 (I'm no longer sure aboute the "1991" part, it could have been "1992" or "1993", don't care).

Of course this ended up in their library, along with a matching card in the index (they still had a paper-based index back then).

I actually got asked by someone if he could base his doctoral thesis on my "phenomenal" findings. I told him to do the math, and contact me again if he still considered this a good idea.

Comment Ages ago.... (Score 1) 251

I built a C64 expansion card containing 256KB of EPROM and 256KB or RAM that I could use via bank switching. As I had no fancy layout tool back then, I had to draw the layout in a paint program (taking into account that the nine needle dot matrix printer had a 216x256 raster!), matching both sides manually, print it, find a photocopier that actually reduced the size by 50% without bending it totally out of shape (I learned the hard way that photocopiers back then had the habit of being a bit fish.eyed when it comes to resizing), make the PCB, and drill a gazillion holes with a hand-kranked drill. Most vias were placed wherver there were wired elements or sockets, but quite a few vias had to be made by soldering a bit of wire on both sides. A horrible hack job in retrospective, but it worked flawlessly from the beginning!

Comment Strange relation (Score 1) 106

The VW (and probably others, I don't believe that only VW cheated - What miracle did they all work in unison to be 30x better with emissions than VW?) problem is the engine, not the general "Car Intelligence". I believe that the VW scandal will lead to more electric cars in the future (not electic replacing diesel, but a shift where gasoline enters the diesel domain, while at the other end electric engines cut their margin of the gasoline market).

What will happen in the future, though, is that the certification authorities will want to see, examine and understand the source code. Which will not only prolong the certification, but also make it way more expensive. It will also force the car manufacturers to cleanly separate the engine control domain from the other control domains in the car, so they can limit the skope of openness to this one domain only.

Slashdot Top Deals

The bigger the theory the better.

Working...