Freesteel Blog » 2006 » May
Tuesday, May 30th, 2006 at 1:16 pm - Machining
I’m always losing track of time. I expect the days on either side of it may collect some stuff to do en route through the South East.
We leave Harwich on 14 June at 18:00.
We return to Harwich on 24 June at 11:30.
No, we don’t use those awful airplane things if we can help it.
Go by ferry instead, and you’ll discover how many essential journeys can be replaced with use of the telephone. Lots of time to do programming type stuff without the distraction of the internet. Gotta be good.
Wednesday, May 24th, 2006 at 10:22 am - Whipping
I’ve been mentioned by name in an article about the Statute Law Database in The Times of London yesterday.
The latest in a series of missed deadlines for public access to statute law is now set for September. “The delay is unnecessary,” says Julian Todd, an IT developer and co-creator of the civic websites publicwhip.org and theyworkforyou.com. The sites mine Hansard to provide an easy way to view an MP’s voting or attendance record and keep track of debates or issues in Parliament. “I can’t comprehend what the DCA thinks it is gaining by not giving us a database dump of the law.”
Martin just got an email back from Austria, a year after he sent them an inquiry about it. Click on http://www.high-speed-machining.at/award/award.html to check out the competition.
If, like me, you don’t know German, babelfish still works well (paste in the URL and select [German -> english], and surf by proxy).
This easy-to-enter competition (if you have access to a machine tool) will send you a simple part of a chocolate rabbit mold, and you have to send them videos and a tape file of your work by September it seems. They specify the metal type, and there is a reward for innovation, not just for brute speed. Any company who enters the competition and loses will say the results are completely unimportant. If they win, it will be all over their website. Often awards go to bigger companies because it makes much better publicity.
I’ve often said we need a Which? guide to CAM software in Europe, and that the EU could make a greater contribution to the manufacturing base by funding a single workshop whose full-time job it was to cut parts using different CAM systems, and post up honest reports about them on the web, than by paying for dodgy engineering startups no longer in existence.
Clearly this would not be in the interests of large CAM software vendors, particularly those who want to make as much money as possible with whatever software they happen to own, but it would make a world of difference for customers and developers who form the basis of the economy to know what’s good, what’s not, and what is it they should be doing. Unfortunately, since “business leaders”, such as CAM software sellers, are more able to get the ear of government ministers than people representing the interests of the users, things get tilted unfairly in their direction.
So we’re on our own, as we are with the great software patent battle. I wish it was not too much to ask for that the software users could somehow get together and raise a subscription of 100 dollars a year to a fund that pays for this service in the form of one guy with a machine tool working full time on their behalf to become an expert of all CAM systems out there, and be able to answer questions about them. Unfortunately, everyone is competing against everyone else, and if you find a CAM system that works better than your neighbour’s, you don’t want to advertise the fact because it may be giving you the edge in winning contracts against him. And so this propensity gets used against you, while the CAM sellers understand their common interests of locking users up into whatever system they have signed up to, and of faking quality.
A trivial start would be for users to ask their software suppliers to enter this competition, and to tell any approaching salesmen that they are only intersted if their name is on the list of entrants. Winning is not so important as participation, although the fixation on “being the best” according to some narrowly chosen and often irrelevant measure it a major problem. What matters is “being good enough” (such as, within 10% of the best) and of demonstrating the willingness to benchmark how things are, and make sustained measurable improvements all the time.
Mmm. Wobbly pencil milling paths to fix now.
I finally caved in and cleaned up the front page to Freesteel, which advertises this Adaptive Roughing thing, now that we expect to get a whole rush of visitors. It was just a little too yuk for me to handle anymore, especially as I’d been playing around with Publicwhip this morning. I wish we had better videos.
This May has been the rainiest May on record in England, so I’ve not got out very much. Nevertheless, here’s a few pictures from the past couple of months.
I went to this Mach 2006 exhibition in Birmingham on Tuesday, a day early, because I thought Martin was supposed to be back from his hols in Germany. So I wandered around the place on my own with half our small pile of “brochures” (one double-sided page with pretty pictures) aware of just what a stupid idea this was going to be. Normally I come into these shows belonging to a company, so at least I’ve got a home (the company stand) to retreat to when it becomes all too much.
In absence of that, I went straight over to GS Productivity Solutions, run by a couple of people who have posted onto this blog, who are very new Gibbscam resellers. The problem was they were way too busy pitching and selling, and never had any time to talk to people without any money. This Gibbs software must be really hot. There’s hardly any advertising of it in England. Bizarrely, Gibbs can’t machine STL triangle files, and I don’t know how to write algorithms that machine on anything but STL triangle files. Gibbs, like Delcam, runs a proper bulletin board, which is a Good Thing. AFAIK Vero and NC Graphics don’t, and Mastercam has a password protected one I can’t see. Not hosting one in public is a sign of cowardice and ought to mean the same thing to users as not being willing to publish one’s accounts means to investors.
So I was left alone and accidentally walked past the Vero stand. I accepted their dixie cup of coffee, and from that moment on had a terrible, eyewateringly bad headache until the next morning. It hurt so much I didn’t care about anything, and no longer gave a fig about making a fool of myself trying to tell people about my new algorithm for toolpaths.
I pestered six companies, all of them either machine tool manufacturers, or cutter makers. Although they don’t have any software for which they could buy our algorithm to use as part therein, I said that if they liked it they could tell other people that they should want to buy it, so they could tell the third group of people — the CAM software companies — to buy it from us so they could then buy what the machine tool maker recommended them to buy, and therefore something would get done. I don’t know how often you have to rely on a double-indirection like this in marketing, but those who saw the point of the algorithm were usually sympathetic.
Then I caught the train down to Bristol for the night and had an interesting conversation with my friends who, in between the chuckling at the idea that I was actually trying to sell something, tried to explain why business requires such a thing as “commercial confidentiality”, even though the convention works totally against the customer’s interests which would be better served by a free-flow of ideas and implementable solutions between all potential suppliers so that the options emerge into the market faster and more evenly, and there is better competition. Apparently, unless companies can keep profitable ideas to themselves to cream off as much market share as possible, no one will bother to invest in any new and risky idea because it’s always more profitable to copy it from the others. Therefore nothing will ever get done by anyone.
As with almost all economic doctrine, it collapses so completely whenever you have something which questions it, you should wonder why anyone bothers to believe it. Is no one old enough to remember the story behind the IBM PC clone, of which we — all of us — have massively benefitted? Do you really think the PC would not have been invented if IBM didn’t believe they could do this to all of their competitors? Innovation happens in spite of today’s business practices, not because of it.
I had to remove a couple of STL files from our collection of uploads. (One was called
Geneva.stl, the other
Suspension Upright.stl). The files were broken: in the header section claiming to be binary files, and indeed containing binary data, but also containing keywords only allowed in ASCII formatted STL files. No idea who uploaded these files, but if the person who uploaded them reads this and wants some more information on what is wrong with these files, please comment here.
Rather than do anything useful with wikifying machining algorithms (someone has to start me off by making the pages), I’ve been getting involved with Wikiscuba, which is far more interesting, as it allows me to write a page on Kayak Diving in the hope that some other buggers out there are going to start doing it, and the fact that I got out to sea for only the second time this year and it was glorious.
Back to machining stuff
I was pleased to find a forum for users of Delcam hosted at the Delcam website, which does it properly. I don’t know how much moderation and censoring goes on with it, but it’s a huge step in the right direction. You get to see gripes, bugs, and work-arounds as the users struggle or succeed with the software. At last a CAM company that is confident enough to encourage users to form a community and do this, and allow members of staff to apparently freely post to it. This is the way forward. Normally, it’s this two-way communication between development and users that the management normally wants to block from happening (or at the very least not encourage), because that’s how they grow their power. You wouldn’t need any management at all if there was no barrier between the users and developers, and programmers just got on with doing stuff that was designed to be used, rather than simply sold.
Ideally users should be talking to one another somewhere that isn’t branded, so if they want to share something that is of great benefit to themselves and to the wider efficiency and productivity of the economy, but slightly detrimental to Delcam’s share value, they can’t be stopped. A place that would almost suffice is: CNC Zone CADCAM area, but even here there are controls. There’s a lot of advertising, and you are forbidden from talking about alternative forums that users could join. Also, Delcam employees aren’t going to be posting to there, so you will be more or less talking among yourselves.
I often keep eyes on Surfware to see how much success they’re having with marketing Truemill and advertising their bogus software patents and funny trademarks. Not much changes from month to month. The cute little dog with a rocket strapped to its back has disappeared. The real information can only be found users discussing that company and its products, half the time hoping that someone in the company might eventually read it and understand what they want.
For eye-candy, I’ve got the Ministerial Whirl going again. Click on the button marked “Today” and skip back and forward a day to inspect how Tony Blair has reshuffled the cabinet. The guesses are that he’s moved the Foreign Secretary Jack Straw out of his job so he can’t go round the world telling everyone that the idea of bombing Iran is “Nuts”. This is so they can scare the American people about the mortal danger from Iran, bomb or threaten to bomb it in self-defence during September and October, and win the mid-term congressional elections in November as a reward for their prowess. What worked in 2002 can work again in 2006, particularly as (a) they have no other options, and (b) they have no scruples whatsoever. They do not care how many bodies they climb over on their way back into office, unless it completely blocks the doorway.
Tuesday, May 2nd, 2006 at 11:39 am - Machining
Okay folks, since no one else is doing it, I’ve started wikipedia-ing on machining topics. See my userpage. Some of this material should be moved to a discussion page elsewhere.
The problems with the terminology are severe. I’m not sure we can call the software we are doing CAM, because that’s a way too broad a category in my opinion, since it must cover all robotics. So, what do we call it? Do I have the list of different machining strategies right? Are there lots of alternative names? What are they? What do we do about the articles G-code and CNC being almost the same? Is G-code the correct term, or is it ISO, or something else? Do we put in a page for “post-processor”?
Once we get somewhere with this I’ll start to be able to elaborate the pencil milling strategy as I’ve implemented it. Someone else can explain BKChoi’s ZMap, which appeared to me when I read it to be similar to cutter location against triangles, except implemented in a utterly daft way. Maybe I have it misunderstood.
People aren’t keeping us busy with bugs and feature requests on the Adaptive Clearing algorithm, probably because they’re not using it. For those who have been emailed copies of the executable, you’re supposed to tell us if it doesn’t work. I spent two years maintaining the post processors in NC Graphics; post processors are never right, especially the one that comes with this.
So, in absence of activity on that front, I have been compelled to write a pencil milling algorithm for Cimco. Since I’m using the same underlying structure as for the const scallop, the subdividing is quite free-form. The one I wrote in Depocam/Machining Strategist is based on quad-trees and is very complex. I remember there being a lot of memory issues, which made the structure even more compressed and hard to program.
No one can tell how this algorithm will perform by comparison to that one in the long term. Like any construct or machine, until it’s been tested, tuned, adjusted, and optimized for circumstances in the real environment, it’s hard to know how far it will go. With a more free-form structure than a quad-tree, there is a lot more room for change. A quad-tree, however, will always be a quad-tree; all that can happen is it gets implemented more efficiently.
No doubt the Danes will start finding problems with it during the next week, before I get stuck into the next algorithm. Did someone mention Rest Area Detection? Oooooh Noooooooooooo!