10:21 <dsheets_> nodes or subgraphs?

10:21 <avsm> just the fig captions

10:21 <avsm> dsheets_: first one could explain what the shapes mean

10:21 <avsm> ideally without disrupting layout too much

10:21 <avsm> fig 4

10:30 <dsheets_> avsm, how about that?

10:30 <avsm> dsheets_: yallop_: reading

10:30 <dsheets_> it seems to reiterate the text, thoguh

10:32 <avsm> yallop_: dsheets_: i think thats fine actually; its a concise summary

10:33 <avsm> wonder if something similar for fig 9 would be useful; its complex

10:33 <dsheets_> ok

10:34 <dsheets_> yallop_, do we mention late binding anywhere? we could instantiate with a module which has a run-time registry of bindings

10:35 <yallop_> I don't think we say anything about that

10:35 <avsm> dsheets_: yallop_: hows my edit to fig 1 caption? still make sense? shorter

10:39 <yallop_> still makes sense to me

10:39 <avsm> layouts been a little changed with this, will tweak after last caption change

10:41 <yallop_> looking at this struct layout detection workflow I'm wondering whether we should have acknowledged Rube Goldberg

10:42 <avsm> it needs Mechanical Turk bindings to automate customer support

10:43 <dsheets_> ok, it's over 13 pages now

10:43 <dsheets_> and there is some col broken code on 9

10:43 <dsheets_> now, to add rows to the table

10:44 <avsm> ohoh

10:44 <avsm> im fixing bibs to trim urls

10:44 <avsm> easily gets us back down

10:44 <avsm> someone fix the col broken code

10:45 <dsheets_> bib counts to limit?

10:46 <yallop_> yep

10:46 <yallop_> I'm wondering whether we should break the sentence up a bit in the first caption

10:46 <yallop_> Struct layout detection: a functor that constructs C type representations is applied to a query generator to generate C code. The C code compiles to an executable oracle that produces an ML module containing layout data. The original functor is applied to this layout module to link the type data into the program.

10:46 <yallop_> better/worse?

10:47 <dsheets_> better, go for it

10:47 <avsm> better

10:47 <yallop_> a little less "house that jack built", I think, but I don't know whether that's an improvement

10:47 <dsheets_> look at 9, too

10:47 <avsm> back to 12 pgs

10:48 <avsm> saved some more space in acs

10:48 <avsm> acks

10:51 <avsm> ok pushed page 6 overflow fix

10:53 <dsheets_> more users in table

10:54 <avsm> uploaded new snap to hotcrp

10:55 <yallop> pushed tweaked text and layout for figure 9

10:55 <yallop> now we have some spare space again

10:56 <avsm> yallop: woah no

10:56 <avsm> that messes up

10:56 <yallop> ok, shall I revert?

10:56 <avsm> yep

10:57 <avsm> just check last version i uploaded to hotcrp

10:57 <avsm> as of commit d7644dd99d118b16ccc877d2748d65389407155e

10:57 <avsm> Author: David Sheets <sheets@alum.mit.edu>

10:57 <avsm> Date: Sat Feb 28 10:52:52 2015 +0000

10:57 <avsm> cmeleon: some more users

10:57 <dsheets_> i put a \newpage before 5.2 and it fixed some awkward sections

10:57 <dsheets_> (not committed)

10:58 <dsheets_> section titles lined up with page/col breaks better

10:58 <dsheets_> but a hole at end of p9

10:59 <avsm> leave submitted version i think

10:59 <avsm> nothing terribly wrong

10:59 <yallop> ok

10:59 <dsheets_> ok, ttyl

10:59 <yallop> reverted that last change in any case

11:01 <avsm> nice work!