#linuxcnc-devel | Logs for 2013-10-08

[12:39:29] <mhaberler> here's my paper for the OSADL conference - I'd welcome any comment, but need to send it off in 3hrs: http://static.mah.priv.at/public/paper.pdf
[12:43:58] <mhaberler> bbl
[12:46:57] <archivist> skim read it and did not find any obvious problems
[12:48:05] <cradek> mhaberler: you should be using `` and '' for quotes instead of "
[12:49:03] <archivist> I didnt worry about stuff like that, facts more important
[12:49:11] <cradek> you've misspelled daemon as demon (not just a UKism - they have different meanings to americans too)
[12:49:35] <cradek> (in Fig 4)
[12:50:52] <archivist> section 6 page 5 "than assumed time; or modify" should that be a , not ;
[12:51:51] <cradek> probably
[12:53:22] <archivist> the diagram at the top left of section 1.1 could be larger to be able to read the text
[12:54:13] <archivist> I have to look close with the viewer set to 200% to read the small print
[12:54:16] <cradek> use of a space before uS is inconsistent
[12:54:46] <cradek> the blue "read" in the hal commands is weird
[12:56:37] <cradek> #6 and #8 in the bibliography are kind of mangled
[12:58:36] <archivist> expand fig 1 so all the bib is on one page and go to single column on the bib to fix that
[12:58:55] <cradek> yeah, not sure if Fig 1 would be readable even on paper
[12:59:47] <archivist> needs to be at least double, there is space on the last page for expansion of it
[13:00:36] <cradek> if the bib could be in one column it would help unmangle the URLs (maybe put a linebreak after the URL)
[13:00:52] <cradek> depends how much of the formatting is dictated
[13:01:58] <cradek> I'd prefer GUIs to GUI's but people don't all agree about this
[13:02:14] <archivist> I think the added space and hyphen are due to the format
[13:02:52] <archivist> bib needs to be left justified
[13:03:33] <cradek> in Abstract: present an overview OF
[13:05:46] <cradek> mhaberler: if you want me (native-english-speaker-latex-user) to go through this I'd be happy to, send me the source, this is pretty inefficient
[13:34:39] <mhaberler> baack, well thanks for reading this
[13:34:48] <mhaberler> cradek: I'll make a repo
[13:35:13] <mhaberler> really appreciated - latex 'deer in the headlight'
[13:35:45] <cradek> heh how much time is left?
[13:35:45] <archivist> I am at the latex noob stage too
[13:38:36] <mhaberler> formally 2.4hs, but it wont be an issue I think
[13:46:33] <mhaberler> ok, this should work: git clone git://git.mah.priv.at/osadl-paper.git
[13:47:19] <cradek> yes and it even builds, yay
[13:49:45] <mhaberler> ha
[13:50:51] <cradek> you'll have to fix demon->daemon uflavor
[13:50:59] <cradek> er, you'll have to fix demon->daemon in uflavor
[13:51:34] <mhaberler> ok
[13:54:41] <mhaberler> pushed
[13:58:00] <mhaberler> i could rearrange fig1 so its more high than wide
[13:58:22] <mhaberler> it comes out too small
[13:58:38] <mhaberler> or if you have a double-column trick up your sleeve - fine
[13:58:43] <cradek> that would be a good way if you want to keep the columns
[13:58:51] <mhaberler> ok
[13:58:56] <cradek> I wasn't going to touch that, working on language/typography so far
[14:01:18] <mhaberler> duh. I used http://www.digikey.com/schemeit and that figure wasnt saved.
[14:01:28] <mhaberler> not an opton anymore
[14:01:32] <cradek> drat
[14:01:48] <mhaberler> it's ok - people are suppose to read online and use zoom
[14:26:51] <cradek> mhaberler: do you prefer american or british? I see behaviour but flavor.
[14:27:18] <mhaberler> american, never lived in blightey
[14:27:48] <mhaberler> reading up on demon vs daemon..
[14:28:34] <cradek> to make it harder, consider the bsd daemon (their old mascot) is a demon
[14:28:57] <cradek> (but I am sure you should use daemon)
[14:29:25] <mhaberler> seems 'demon' means 'devil' more directly, but dict.leo.org is fuzzy
[14:37:06] <cradek> I don't understand how/why latency causes position hunting on a still machine
[14:38:21] <mhaberler> thats from Charles, need to ask him
[14:39:24] <zultron> daemon!
[14:39:56] <zultron> (drive by comment while getting an afternoon coffee ;)
[14:40:56] <skunkworks> went to school with a damian
[14:53:15] <cradek> mhaberler: http://timeguy.com/cradek-files/emc/0001-Language-and-typesetting-cleanups.patch
[14:53:30] <cradek> mhaberler: I've never seen a binary patch before...
[14:53:47] <mhaberler> well thanks at lot!
[14:53:49] <cradek> good stuff is at the end
[14:53:53] <mhaberler> that was very nice of you
[14:53:58] <cradek> welcome
[14:54:07] <cradek> I hope your talk goes well
[14:54:58] <cradek> I'm not 100% happy with it but I'm out of time I can spend on it
[14:55:00] <cradek> brb
[14:57:30] <mhaberler> thats just fine
[14:57:42] <mhaberler> Architecure.. I like it, in retrospect
[15:01:48] <cradek> I failed at unsplitting Figure 3 and its text
[15:01:49] <mhaberler> I took the liberty to add you to the authors
[15:02:12] <cradek> aha, maybe that extra text fixed figure 3 :-)
[15:03:05] <mhaberler> positvely, this must have been it
[15:03:12] <mhaberler> anyway, off it goes
[15:03:34] <mhaberler> 2:10 before deadline, actually I could have a beer first
[15:06:30] <cradek> heh my lshort2e.dvi is dated 1995
[15:06:49] <cradek> I bet there are updates since then...
[15:17:47] <mhaberler> submitted - done! thanks a lot again!
[15:17:56] <cradek> yay, welcome
[20:17:52] <memleak> gabewillen, hello!
[21:57:34] <kwallace1> It looks like a work offset table lives in interp_convert (or something similar) and there is a work offset table in emc.var. When LinuxCNC loads emc.var gets loaded into the interp table. Then I could change the interp table with MDI or run a G-code file, or I could edit emc.var. So it seems the two tables can be the same or different at different times.
[22:00:30] <cradek> yes if you edit emc.var yourself, you're going to get bitten.
[22:00:51] <kwallace1> I have always just used the touch off feature and looked at emc.var to verify my changes, so I haven't had too much trouble with work offsets, but using G10 seems to be able to put a wrench in the works.
[22:00:56] <cradek> the purpose of emc.var is to preserve interpreter state from one run to another, nothing else
[22:01:10] <cradek> touch off calls exactly G10
[22:02:02] <kwallace1> So I should pretend that emc.var doesn't exist?
[22:02:37] <cradek> as a user, postively yes you should ignore it
[22:04:29] <cradek> ALSO as a gui author
[22:04:50] <cradek> only if you are working inside the interpreter guts should you be interested in it
[22:07:52] <kwallace1> I'm still trying to figure out what a work offset utility should look like.
[22:08:43] <cradek> I think it would be very hard to make a gui that does everything that G10 lets you do
[22:09:41] <cradek> I mean intuitively - obviously you can make a gui that lets/helps the user make G10 commands, which is what touchy's set offset/set tool buttons do
[22:10:25] <kwallace1> My next pass might be to have an empty table that allows the user to enter values, then have a load button for each entry.
[22:10:29] <cradek> to me, a table full of numbers is a solution looking for a problem, and is not a task-driven approach
[22:11:05] <cradek> what problem are you solving? what is your user doing when you think this is a useful tool?
[22:12:14] <cradek> if your coordinate system is rotated, g10 lets you set pairs of numbers in ways that would be difficult to calculate and enter as raw numbers in the varfile-like table
[22:12:35] <kwallace1> Part of the problem is figuring out what the problem is.
[22:12:59] <cradek> haha, as jwz famously said, "now you have two problems"
[22:14:38] <cradek> I'm puzzled if you've decided on a UI but don't yet know the task someone wants to accomplish by using it
[22:14:51] <kwallace1> It started out easily enough, "Let's put a work offset table next to the tool table", "Yah, that should be easy."
[22:15:57] <cradek> in most situations g10 is the best way to "edit" the tool table as well [the only exception is adding a new tool, where I think g10 can't do it]
[22:16:46] <cradek> rotate your coordinate system 13 degrees and try to set meaningful tool offsets without using g10 or touch off, and you'll see what I mean
[22:16:59] <kwallace1> But using G10 from the command line requires some abstract thinking.
[22:17:23] <cradek> yes, something like touch off can usefully hide some of that
[22:18:05] <cradek> you work too late for me, I'm off to bed again, cheers
[22:18:38] <kwallace1> Thank you again,