| --- Log opened Thu Dec 19 00:00:19 2019 |
| 00:38 | | Kindamoody is now known as Kindamoody[zZz] |
| 00:59 | <@Reiv> | McMartin: Yes, that was exactly the point he was making |
| 00:59 | <@Reiv> | It felt like he was making the point mostly to note "This is not a How To Do These Things, it is a How To Think About These Sort Of Things, do not mix them up and just use goddamn libraries it's fine kids" |
| 01:00 | <&McMartin> | Yeah |
| 01:00 | <&McMartin> | That said |
| 01:00 | <&McMartin> | Some of the extremely basic cases are handy to be able to throw together if your language's libraries are garbo |
| 01:00 | <@Reiv> | One imagines he once marked an assignment where a student dutifully re-implemented /everything/ while working on his Actual Assignment and inspired a hearty sigh at Missing The Point Entirely. |
| 01:01 | <&McMartin> | It really is much easier to use "intrusive pointers" to allow your custom data structure to behave like a linked list |
| 01:01 | <@Reiv> | Oh, sure, that's true too, but not why you'd go through /all/ the encryptions instead of just The Best Half Dozen |
| 01:01 | <&McMartin> | Than it is to have a properly generic linked list structure that you then bash into shape and compliance with everything else |
| 01:01 | <&McMartin> | (Less true when the language's core library includes them and everyone standardizes on them, but when they don't...) |
| 01:02 | | * Reiv nods |
| 01:04 | <~Vorntastic> | My experience with vornonoi soured me heavily on linked structures |
| 01:05 | <~Vorntastic> | Like it looks pretty. At first. |
| 01:05 | <&McMartin> | Mmhmm |
| 01:05 | <&McMartin> | I mean, if i'm in a bad mood |
| 01:06 | <&McMartin> | One nice thing about graph structures being hard to represent is that then you don't *use* them |
| 01:06 | <&McMartin> | And then everybody wins |
| 01:07 | <@Reiv> | Linked structures? |
| 01:07 | <@Reiv> | Why is linked structures bad |
| 01:07 | <~Vorntastic> | So |
| 01:07 | <@Reiv> | I mean, I know what a linked list and/or set of objects /is/ but I'm curious why is /bad/ |
| 01:09 | <~Vorntastic> | This might take a minute to explain so |
| 01:17 | <~Vorntastic> | Vornonoi uses half edge data structures. Each one is attached to its twin (same edge, but other face attached) and next and prev (same face but cycling around it) |
| 01:17 | <~Vorntastic> | Of note, this is doubly linked in all three ways |
| 01:19 | <~Vorntastic> | In linked structures, it's convenient: I refer to an edge the same way I examine its structure |
| 01:20 | <~Vorntastic> | And when doing things "locally" that's a win, because it's simpler to manipulate a few things |
| 01:20 | <~Vorntastic> | But |
| 01:23 | <~Vorntastic> | Not everything is local |
| 01:24 | | Degi [Degi@Nightstar-pmefj5.dyn.telefonica.de] has quit [Ping timeout: 121 seconds] |
| 01:25 | | Degi [Degi@Nightstar-jcpjli.dyn.telefonica.de] has joined #code |
| 01:28 | <~Vorntastic> | The dual operation - converting "roads between capitals" of delaunay into "borders between regions" of voronoi |
| 01:29 | <~Vorntastic> | Is global: all edges must be manipulated, and thus collected |
| 01:30 | <@sshine> | I wonder if there's a way to reverse engineer what3words. |
| 01:30 | <@sshine> | I'd like to see if there's a square nearby my area that ends with a word that is also a top-level domain, e.g. 'cloud' |
| 01:31 | <~Vorntastic> | So you have to - because it's not a linear linked structure - perform search-like traversal to get them all |
| 01:33 | <@sshine> | :( |
| 01:34 | <@sshine> | I just hit up some random combination and found some pretty trippy art: https://what3words.com/happy.forest.cloud |
| 01:35 | <~Vorntastic> | Worse, you also may need to do this for structure deletion: refcounting fails on doubly-linked structures, because every reference is circular. As the structure programmer you must write a deletion routine if you want it to not stick around for gc to run, and in many languages you need to insist that it gets run |
| 01:36 | <~Vorntastic> | This deletion problem is why I switched to array structures and indexes into those arrays instead of references. Without explicit deletion, my demo program was literally running out of memory, and when I switched to arrays it stayed under 10mb |
| 01:39 | <~Vorntastic> | In addition, array structures I could literally copy into Excel and examine directly in cases of failure, and I can also serialize them trivially in json |
| 01:43 | <~Vorntastic> | Also insert discussion of memory continuity and processor caching here |
| 01:46 | <~Vorntastic> | Catalyst can tell you more on these subjects |
| 01:58 | <@Reiv> | Ha, yes indeed, an interesting point |
| 01:58 | <@Reiv> | Cheers, Vorn! |
| 02:32 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 03:51 | <&McMartin> | https://cdn.discordapp.com/attachments/457312182842032140/657027158203564054/kJW8ImA.png |
| 03:52 | | * Alek snerks. |
| 05:10 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has quit [Operation timed out] |
| 05:42 | | Derakon is now known as Derakon[AFK] |
| 05:56 | | celticminstrel [celticminst@Nightstar-2kik29.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 08:39 | | Kindamoody[zZz] is now known as Kindamoody |
| 08:45 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has joined #code |
| 08:45 | | mode/#code [+o himi] by ChanServ |
| 08:50 | <@sshine> | I needed to redo my VPS setups and wanted a domain for the new server, so I got 'eta.solutions'. I figured that might be a pretty cool company name for freelancing through. so I check the national company register, and it turns out a guy called Simon just closed his company, 'Eta Technology Solutions', and the guy studied at my university and lives 500 meters from me... |
| 08:50 | <@sshine> | sorry, his name was Simone. he's italian. |
| 08:52 | <@sshine> | I started googling around for 'Eta Solutions' and realize it's a terribly unoriginal name. there are so many companies called that. so I thought 'Eta Software Solutions'. at least it wasn't 'Eta Technology Solutions', or I might as well have changed name to Simone. :-D |
| 09:15 | <&McMartin> | For those who are curious: https://github.com/michaelcmartin/bumbershoot/blob/master/win32/WinLights.asm and https://github.com/michaelcmartin/bumbershoot/blob/master/win32/WinLights.rc |
| 10:11 | <&McMartin> | Yep. That was probably to avoid the name of the old toy, because doomlawyers |
| 10:11 | <&McMartin> | Oops, way mischanned, sorry |
| 10:14 | <~Vorntastic> | Instead of glory kills they hand out subpoenas and documents come out |
| 10:17 | | * TheWatcher eyes this email |
| 10:19 | <@TheWatcher> | I apparently need to spend some time writing documentation, some madman at another uni wants to use some of the software we use to support teaching |
| 11:51 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 13:15 | | Kindamoody is now known as Kindamoody|afk |
| 17:11 | | Vorntastic [uid293981@Nightstar-2dc.p8m.184.192.IP] has quit [[NS] Quit: Connection closed for inactivity] |
| 18:46 | | Vash [Vash@Nightstar-sjaki9.res.rr.com] has joined #code |
| 19:49 | | Vash [Vash@Nightstar-sjaki9.res.rr.com] has quit [[NS] Quit: Leaving] |
| 20:05 | | Kindamoody|afk is now known as Kindamoody |
| 20:30 | <&[R]> | <mrguyorama> ah, great, nobody knows the process for approving access requests because it changes every three days |
| 20:30 | <&[R]> | <mrguyorama> in the middle of the holidays. Our busy season |
| 22:23 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 121 seconds] |
| --- Log closed Fri Dec 20 00:00:20 2019 |