I've never used it for robotics (used Stamps and 68hc11s before, though), ARM is a pretty sweet architecture. They're low cost, low power, and have really nice assembly code.
Yes, mathematics is full of ambiguities like this. Inverse has two meanings - rational inverse, as it is here f(x) = 1/sqrt(x), or, as you were thinking, functional inverse - f(x) = x^2.
Not really, the submitter did not say they would use the exploit if refused. The suggestion is rather that someone likely will use the exploit. It's not really coercian, intimidation, or a direct threat.
Yet it certainly is not general purpose. It'd be essentially impossible to write a text editor, for example. At least without having the CPU do alot of the work.
Re:It's like nothing we've seen .. since Linux
on
A New Kind of OS
·
· Score: 0
I think programming with boxes and lines is just another way to program. You're right, it doesn't really make it much easier. However, it might be better for more visual people. There is a game called mindrover where you wire together logic components to make a robot's brain. Pretty cool stuff really. Box type data pipelines are ideal for paralel processing. Visual programming gives a much better representation of this than textual.
I think for some difficult programming tasks (namely those which you can automatically ascribe a numerical fitness to a solution), new paradigms will make them near effortless in comparison.
I really like your description of the problems of GUIs. I think a GUI can and will be vastly superior to a command line in nearly all tasks, we just haven't done the right things. I think if the backend is intuitive, the frontend can be elegant.
Also, as far as the Neal Stephenson thing, Apple had a GUI well before Microsoft.
I kind of wish the internet was still limited to the nerds in some way - not that it's less free for everyone to use - it's just that if you want to put something on the internet you would have to seek out a nerd facilitator...
Anyway, yes, most of the internet is now rubbish. Thankfully search engines do a decent job of sorting through this rubbish, but as the rubbish becomes more important than real information that might change. That is, a retarded blog page linked to from tons of spots might rise to the top of a google search, as oppoed to a reasonable, informative resource.
I suppose it might be possible to start again - offer something like the internet yet make it better. I've always wanted a better markup language than html. Screw xml in general - create a new semantic markup language and presentation system. I know that this doesn't actually have much to do with a new internet, however if you're going to have a new network, might as well use a new format as the standard. If people have access to this new network, the browsers will be provided as well. The network could be cobbled together through wireless repeaters, at first a neighborhood of access, eventually a city, etc. Wires between wireless bits could provide fast interconnect between locations. Routing around would likely require complex pathfinding, but after a path is visited it could be optimized. Repeaters could learn where to send packets for fastest delivery. Bridges between the new and the old would also need to be constructed, yet metadata, tagging, commenting, rating, moderating could all be built in.
Most of all with this system is that you wouldn't need fancy dancy servers to produce popular content. The repeaters would have hefty hard drives and work somewhat like bittorrent, extept that everyone actually seeds.
All the kludges of the current internet could be alleviated.
Maybe its a crackpot idea, I haven't thought it out throughly.
Mod parent up!
I was about to say:
There have been robots that can do dead reckoning using steppers and encoders and such for quite a while.
But that is a pretty interesting approach. Way better than the current idea of precision mapping (perhaps using multiple agents). Have a room relationship map rather than a real map.
Subsumption architecture is pretty cool as well. It's rather obvious, but that's probably just because ive known about it for a while.
I think kind of a fuzzy hierarchal subsumption architecture would be better though.
Man, and I thought I was perfectionistic. You could have put all that effort into doing a little coding to make inkscape output what you'd like. It probably wouldn't be that bad. The optimizer someone was working on (to remove unused gradients etc) might have made it into.44, I'm not sure.
Many crashes have been fixed in this version. As for the release notes, I can access them.
Verbatim from the wiki:
Speed
In addition to the Outline mode which makes it much easier to work with complex drawings, this version of Inkscape also provides significant speed improvements in many areas.
Thanks to optimizations in the renderer, Inkscape's screen redraw is faster by at least 10%, and in some cases (such as complex stroked/dashed paths at high zooms) up to three times faster.
Optimizations in the Node tool resulted in noticeable speed gains for node editing. Thus, switching to and from the Node tool (with a path selected), as well as selecting nodes in that tool, are now at least ten times faster than before. Other operations, including curve and node dragging and move/scale/rotate operations on multiple selected nodes, are much faster as well. This is especially important when working with complex paths; with these optimizations, paths containing several thousand nodes, though still slow, are much more usable.
An optimization in the attribute setting method made operations such as moving multiple objects with arrow keys at least 30% faster compared to 0.43. This is especially noticeable when you are moving clones selected together with their original (e.g. a clone tiling), in which case Inkscape now works three to four times faster.
Interface icons are now rendered in the background (from SVG source in share/icons/icons.svg) when Inkscape is idle, rather than waiting for all the icons in a menu to render the first time you pull it up. This eliminates the annoying delay when opening menus for the first time.
Previously, zooming in to view a small portion of a path (especially big and complex path), there was a very noticeable slowdown and memory use increased dramatically. We optimized the renderer to only process the visible part of a path, and as a result the rendering speed is now almost the same at any zoom up to the maximum, providing up to 10-40 times speedup compared to the previous version (the closer is the zoom, the greater is the gain).
The Path > Break Apart command is now dozens of times (up to 100x) faster for complex paths with thousands of subpaths.
bugfixes
Reading a document with an incorrect namespace URI not only did not cause Inkscape to complain, but could also "pollute" Inkscape's internal namespace table, resulting in an "infection" of subsequently saved documents by the incorrect namespace. This is now fixed, but as a result, documents with incorrect namespace URIs will no longer load. You will have to edit them in a text editor to fix the namespaces.
With newer versions of GTK, dragging with graphics tablet pen did not work in some tools and contexts (in particular, in Node and Rectangle tools). This is fixed.
Scaling of objects with stroke in Selector used to cause undesired shifts of the scaled object, as well as scaling it in the dimension which was intended to remain untouched (e.g. slight change in width when you scale only height). All these problems are now fixed, both for interactive scaling by mouse and for numeric scaling via the Controls bar, and for both values of the "Scale stroke with objects" option. Among other things, this means that stroked objects no longer lose snapping on scale, and that the "Default scale origin" option in the Selector tool preferences finally works as designed. Caveat: There may still be problems if you scale a selection that contains objects with different stroke widths.
Scaling of stroke now works for objects that didn't specify stroke-width; before, they always ended up with the default 1px stroke.
The bounding box for text and flowed text objects did not include stroke width. This has been fixed.
Stroke miterlimit on text objects was misinterpreted in absolute units instead of multiplies of stroke width (resulting in miter joins rendered as bevel).
Yeah, places sounds like one of those lame 'Web 2.0' ish names. Under this usage of 'Web 2.0', I'm more of referring to an entire ideology. The simplicity thing started by apple. It seems like it's just itching for an 'i' infront of it to make 'iPlaces'. Ugg. Maybe some people like that kind of thing, like songs overplayed on the radio I get tired of it to the point of hate and nausea.
I can see the 'places' logo now. Under it, it would say "Organize. Revisit. Rethink." or some shit.
Seriously, stop trying to name things with artistic simplicity. Its not working any more.
A new bookmark system doesn't need a name.
Graphics, and physics - meqon kicks ass!
I saw a demo of it, best physics engine ive seen. I suppose the ageia stuff looks very nice, but ive only seen videos and not demos.
I read to the reference of wolfram and stopped.
Sure, wolfram is a smart guy, mathematica is pretty cool. His book is totally conceited and most of its content is unoriginal, without reference to the true creators. In no way did he invent CAs, or as far as i know really contribute much to the field. Conway's game of life was the first CA, I believe. Way before wolrram's time.
Anyone that gives this guy much credit for these things immediatly loses large portions of any respect I have for them.
It's too bad that often times when you find something that might suit your purposes it is badly designed and lacks features. So you end up reinventing it because they did not implement it well.
Yep, just 3.2 grams of pure caffeine administered intravaneously has a good chance at killing you. (this is the lowest reported lethal dose) 10 grams though - man, your dead.
Yeah, I always thought it was smarter to have humans control the robots, rather than cooking up some AI. Sure, AI is cool, but for this application it is really unecessary.
Plus this would give all of this generation's video game addicts a decent job.
I think you do not understand the concept of science. The shows follow the base scientific method you learn in elementary school.
As to the engineering comment, I take it you haven't watched the rainwater-pipe runoff episode, or the one where they disprove the myth of slingshotting immigrants over the border.
Slashdot Mirror
I've never used it for robotics (used Stamps and 68hc11s before, though), ARM is a pretty sweet architecture. They're low cost, low power, and have really nice assembly code.
Too bad that all of those ideas came from outside the company.
Yes, mathematics is full of ambiguities like this. Inverse has two meanings - rational inverse, as it is here f(x) = 1/sqrt(x), or, as you were thinking, functional inverse - f(x) = x^2.
You should read ishmael.
Not really, the submitter did not say they would use the exploit if refused. The suggestion is rather that someone likely will use the exploit. It's not really coercian, intimidation, or a direct threat.
Yet it certainly is not general purpose. It'd be essentially impossible to write a text editor, for example. At least without having the CPU do alot of the work.
I for one welcome our new mouse overloards.
I think programming with boxes and lines is just another way to program. You're right, it doesn't really make it much easier. However, it might be better for more visual people. There is a game called mindrover where you wire together logic components to make a robot's brain. Pretty cool stuff really. Box type data pipelines are ideal for paralel processing. Visual programming gives a much better representation of this than textual. I think for some difficult programming tasks (namely those which you can automatically ascribe a numerical fitness to a solution), new paradigms will make them near effortless in comparison. I really like your description of the problems of GUIs. I think a GUI can and will be vastly superior to a command line in nearly all tasks, we just haven't done the right things. I think if the backend is intuitive, the frontend can be elegant. Also, as far as the Neal Stephenson thing, Apple had a GUI well before Microsoft.
Least its not apple, or else you'd be paying for the update
I've got a bad habit of trying to redesign old technologies.
I kind of wish the internet was still limited to the nerds in some way - not that it's less free for everyone to use - it's just that if you want to put something on the internet you would have to seek out a nerd facilitator... Anyway, yes, most of the internet is now rubbish. Thankfully search engines do a decent job of sorting through this rubbish, but as the rubbish becomes more important than real information that might change. That is, a retarded blog page linked to from tons of spots might rise to the top of a google search, as oppoed to a reasonable, informative resource. I suppose it might be possible to start again - offer something like the internet yet make it better. I've always wanted a better markup language than html. Screw xml in general - create a new semantic markup language and presentation system. I know that this doesn't actually have much to do with a new internet, however if you're going to have a new network, might as well use a new format as the standard. If people have access to this new network, the browsers will be provided as well. The network could be cobbled together through wireless repeaters, at first a neighborhood of access, eventually a city, etc. Wires between wireless bits could provide fast interconnect between locations. Routing around would likely require complex pathfinding, but after a path is visited it could be optimized. Repeaters could learn where to send packets for fastest delivery. Bridges between the new and the old would also need to be constructed, yet metadata, tagging, commenting, rating, moderating could all be built in. Most of all with this system is that you wouldn't need fancy dancy servers to produce popular content. The repeaters would have hefty hard drives and work somewhat like bittorrent, extept that everyone actually seeds. All the kludges of the current internet could be alleviated. Maybe its a crackpot idea, I haven't thought it out throughly. Mod parent up!
I was about to say: There have been robots that can do dead reckoning using steppers and encoders and such for quite a while. But that is a pretty interesting approach. Way better than the current idea of precision mapping (perhaps using multiple agents). Have a room relationship map rather than a real map. Subsumption architecture is pretty cool as well. It's rather obvious, but that's probably just because ive known about it for a while. I think kind of a fuzzy hierarchal subsumption architecture would be better though.
Man, and I thought I was perfectionistic. You could have put all that effort into doing a little coding to make inkscape output what you'd like. It probably wouldn't be that bad. The optimizer someone was working on (to remove unused gradients etc) might have made it into .44, I'm not sure.
Verbatim from the wiki:
Speed
In addition to the Outline mode which makes it much easier to work with complex drawings, this version of Inkscape also provides significant speed improvements in many areas.
bugfixes
This was a very early slashdot, the site was still being updated, in fact. Must have been a dev or hardcore user that reported it.
IIRC they plan to have a foot power thing external to the laptop.
Yeah, places sounds like one of those lame 'Web 2.0' ish names. Under this usage of 'Web 2.0', I'm more of referring to an entire ideology. The simplicity thing started by apple. It seems like it's just itching for an 'i' infront of it to make 'iPlaces'. Ugg. Maybe some people like that kind of thing, like songs overplayed on the radio I get tired of it to the point of hate and nausea. I can see the 'places' logo now. Under it, it would say "Organize. Revisit. Rethink." or some shit. Seriously, stop trying to name things with artistic simplicity. Its not working any more. A new bookmark system doesn't need a name.
Graphics, and physics - meqon kicks ass! I saw a demo of it, best physics engine ive seen. I suppose the ageia stuff looks very nice, but ive only seen videos and not demos.
but the code, is beutiful.
I read to the reference of wolfram and stopped. Sure, wolfram is a smart guy, mathematica is pretty cool. His book is totally conceited and most of its content is unoriginal, without reference to the true creators. In no way did he invent CAs, or as far as i know really contribute much to the field. Conway's game of life was the first CA, I believe. Way before wolrram's time. Anyone that gives this guy much credit for these things immediatly loses large portions of any respect I have for them.
I think we all know who is the 'fanboi' here.
It's too bad that often times when you find something that might suit your purposes it is badly designed and lacks features. So you end up reinventing it because they did not implement it well.
Yep, just 3.2 grams of pure caffeine administered intravaneously has a good chance at killing you. (this is the lowest reported lethal dose) 10 grams though - man, your dead.
Yeah, I always thought it was smarter to have humans control the robots, rather than cooking up some AI. Sure, AI is cool, but for this application it is really unecessary. Plus this would give all of this generation's video game addicts a decent job.
I think you do not understand the concept of science. The shows follow the base scientific method you learn in elementary school. As to the engineering comment, I take it you haven't watched the rainwater-pipe runoff episode, or the one where they disprove the myth of slingshotting immigrants over the border.