IRC logs for #openttd on OFTC at 2023-11-12
⏴ go to previous day
00:16:52 *** bryjen has quit IRC (Quit: Leaving)
00:39:05 *** ChanServ sets mode: +v tokai
01:00:04 *** _aD has quit IRC (Ping timeout: 480 seconds)
01:34:15 *** WormnestAndroid has quit IRC (Remote host closed the connection)
01:34:22 *** WormnestAndroid has joined #openttd
02:13:48 *** Smedles has quit IRC (Read error: No route to host)
02:14:25 *** Smedles has joined #openttd
02:51:10 *** Wormnest has quit IRC (Quit: Leaving)
02:54:26 *** Smedles has quit IRC (Quit: No Ping reply in 180 seconds.)
02:54:27 *** Smedles has joined #openttd
03:03:50 <bungus> big balls to comment on an issue thread with the same username after getting banned for doing so in the past lol
03:24:31 *** Smedles has quit IRC (Ping timeout: 480 seconds)
03:32:50 *** D-HUND has quit IRC (Ping timeout: 480 seconds)
04:08:06 *** Smedles has joined #openttd
04:08:42 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
04:12:18 *** Smedles has joined #openttd
06:47:55 *** Smedles has quit IRC (Quit: No Ping reply in 180 seconds.)
06:47:56 *** Smedles has joined #openttd
07:49:20 <peter1138> Maybe I should try a different approach.
08:55:37 *** HerzogDeXtEr has joined #openttd
09:24:39 <locosage> peter1138[d]: nice, though I'd probably put
09:25:32 <locosage> also, as much as I'd like to see icons in many more places icons themselves are very bad unfortunately
09:25:55 <locosage> they're too small and indistinct
09:27:35 <locosage> I ended up doing it like this
09:28:16 *** alfagamma7 has joined #openttd
09:31:10 <_zephyris> locosage: Is the better solution larger icons?
09:32:24 <locosage> idk how would that work with original icons though
09:32:39 <locosage> also grfs have their own icons
10:57:19 <michi_cc[d]> Some of the details on how it works are a bit iffy, and the GUI of that versions still needs a lot of love (like scrollbars), but it is cargodest.
11:01:55 <_zephyris> Nice. Will be fun to play. It feels like there's three levels of complexity here, but each with their own merits. Simple: Point to point - distribution - destination :Complex
11:02:43 <michi_cc[d]> My current "in-very-slow-development" branch has fixed some of the edges, but right now it is not in any playable state.
12:20:33 <rau117> _zephyris: Icons without text, no matter how large they are, will still convey less information than small icons with text. Especially if the icons have pixelated style.
12:29:33 <rau117> peter1138: If so, just necessarily make this drop-down list NOT close automatically.
13:09:56 <peter1138> Yes, that's the plan. I tried faking it by reopening immediately, but that has issues, so I did it properly.
13:11:15 <peter1138> _zephyris, maybe not larger icons, but doing 2x zoom icons should allow more detail
13:12:11 <peter1138> The icon size is kind of optimal for where they are used (and where I am using them)
13:23:47 <rau117> Speaking of icons. Maybe they can be added for the industry window?
13:23:47 <rau117> As dp said, before the text, not after. Then it will look neat and clear.
13:27:32 <peter1138> ^ This is why I put them after.
13:27:49 <peter1138> Even the base icons are not all the same width.
13:29:22 <peter1138> (Also I need to fix the divider :D)
13:31:17 <locosage> peter1138[d]: try putting checkmark on the right if you can
13:34:28 <locosage> icons having different width doesn't help either though...
13:40:25 <rau117> peter1138[d]: Or maybe a compromise can be made?
13:40:25 <rau117> Instead of the drop-down list, leave those buttons, but divide them into several rows and use the full name with an icon instead of 2-4 letter abbreviations? Then the interface will remain familiar and fast, and you won’t have to fix the closing drop-down list.
13:40:39 <peter1138> Have you ever played FIRS?
13:41:20 <peter1138> (The closing dropdown list is solved, doing it properly was not much more work than trying to cheese it)
13:42:05 <peter1138> Multiple rows of boxes doesn't work, there's still too many of them.
13:42:29 <locosage> peter1138: ideally icon size should much font size imo
13:42:46 <locosage> kinda hard to do that in openttd context though
13:43:05 <locosage> yeah until you change the font
13:44:37 <peter1138> If you change the font so much that that matters, then the rest of the UI is going to look like arse anyway.
13:47:34 <locosage> not really, I'm using something like 1.5x font size
13:49:16 <locosage> and having 15 pixels for icons instead of 10 would make some difference
13:50:19 <rau117> Or here's another option. As in cargo payment rates.
13:50:19 <rau117> *(yes, yes, I see that drop-down list is ± ready, but maybe someone will like this option better)*
13:53:46 <xarick> Hello. I got a tech question. How much is 1 KB/s in mbps? Speaking about network speeds as reported by HWiNFO64
13:54:46 <xarick> it lists 2 types of kilobits and kilobytes
13:55:06 <xarick> I'm not sure which ones are
13:55:27 <peter1138> One is counted in 1000s, one is counted in 1024s.
13:55:36 <_glx_> * 1000 from KB to B, then * 8 from B to b
13:56:21 <_glx_> and finally / 1000000 for mb
13:58:19 <Rubidium> it's * 1000 for b to mb :D
13:58:24 <xarick> the ISP providers say mpbs or Mbps
13:59:10 <_glx_> in FAI world mbps and Mbps are the same
13:59:56 <_glx_> in real world they of course are not
14:01:23 <peter1138> Mbps is megabits per second. mbps is millibits per second, which isn't a practical unit, so they mean Mbps either way.
14:03:17 <Rubidium> peter1138: it's very useful! Just write in the contract that you guarantee at least 10 mbps for 99% of the time, but you limit the speed to 100 Mbps. If you then provide 5 Mbps for a prolonged time, you're still well in the clear :(
14:13:25 <xarick> I just wanted to make sure whether my ISP is stealing me
14:14:39 <_glx_> if you are on wifi you'll never reach ISP promised value
14:16:06 <Eddi|zuHause> some ISPs have been artificially boosting speed, if they detect a speed test being run
14:19:38 <ahyangyi> Now, if we can shape any traffic to look like a speed test...
14:20:48 <Eddi|zuHause> this was some years ago, but the detection was very crude, like "has speedtest in the URL"
14:22:51 <peter1138> My ISP does not snoop any of my traffic.
14:24:28 <rau117> And there are also different prefixes. Kilo/mega/giga/tera and Kibi/mebi/gibi/tibi.
14:24:28 <rau117> Kilo [...] is used in SI – multiplier 1000
14:24:28 <rau117> Kibi [...] __should__ be used in the IT field, multiplier 1024
14:24:28 <rau117> *But basically everyone ignores this difference and uses kilo [...], because it’s more common.*
14:24:30 <rau117> Greedy manufacturers love to play on the ignorance of users.
14:24:30 <rau117> You can often notice such a difference that the HDD manufacturer claims a value of 1 TB, but in the computer it turns out that this is only 931 real GiB.
14:24:32 <rau117> Because they use SI prefix – 1 trillion bytes, and the real TiB takes up 1 099 511 627 776 bytes.
14:27:46 <ahyangyi> Eddi|zuHause: I wonder if that makes `www.example.com/?from=speedtest` faster than `www.example.com/?from=google` 😛
14:32:06 <tesimk> hello, I was wondeing why 3x font scaling is not supported
14:33:16 <peter1138> Our sprite scaling only works with powers-of-two.
14:33:32 <Eddi|zuHause> rau117: for each step of K,G,T,P, ... it adds a 2.4% difference
14:33:37 <peter1138> But you can use a truetype font instead which allows variable scaling.
14:35:09 <Eddi|zuHause> rau117: it's like disguised price increases by making the packaging unit smaller
14:35:35 <peter1138> But not really because the price of storage is always coming down.
14:35:37 <tesimk> it seems to me that any integer scaling could work just fine for the original font
14:35:39 <peter1138> It's not like food :D
14:36:08 <tesimk> but i haven't found the code responsible for it
14:37:35 <peter1138> The main issue is sprite scaling is linked to viewport zoom levels, and that was designed with 1x, 2x, 4x, 8x etc in mind.
14:38:21 <peter1138> So the zoom level is a bit shift. 1U << 0 = 1x, 1U << 1 = 2x, 1U << 2 = 4x, 1U << 3 = 8x.
14:38:43 <Eddi|zuHause> GUI scaling would work better in steps of root(2), but that means really awkward scaling
14:39:17 <peter1138> This level of scaling trickles through into the sprite caching system, and everywhere.
14:39:31 <Eddi|zuHause> (this is the same scaling as paper sizes)
14:39:39 <peter1138> So yes, while "just allow 3x" sounds feasible, it's a massive job to actually allow in practice.
14:40:01 <ahyangyi> And I think using TrueType fonts is a good enough workaround
14:40:10 <peter1138> And then, when you consider with truetype fonts you can have any scaling you want, it's really very low on any agenda.
14:40:55 <peter1138> The NewGRF data format also only supports 1x, 2x, 4x, etc.
14:40:58 <locosage> having 3x for gui sprites won't hurt though
14:41:01 <peter1138> So there's no way to draw a 3x sprites.
14:41:44 <tesimk> peter1138: i expected there could be such architectural reason for this limitation
14:41:56 <peter1138> Extending the NewGRF data format to allow intermediate sprite zooms is not impossible, but also a lot of work.
14:42:30 <ahyangyi> What about 8x sprites though 😛
14:42:33 <Eddi|zuHause> just reimplement the game with 3D models and 4k textures :p
14:44:17 <peter1138> (Of course, I do have a patch that does runtime scaling of sprites, but 1) it performs badly because it bypasses the prescaling system 2) it does not work on all blitters 3) it does of course suck for non-integer scaling)
14:47:16 <jfs> Eddi|zuHause: ah yes, we can ship the game with an ML model to convert spritesets to textured 3D models! that way every NewGRF will still be playable too
14:47:44 <ahyangyi> everything is NeRF'd and nerfed
14:48:45 <Eddi|zuHause> why does NeRF sound like a crypto scam?
14:49:29 <ahyangyi> 🤔 perhaps because it ends with "F"; sounds financial
15:32:23 *** virtualrandomnumber has joined #openttd
15:33:29 *** virtualrandomnumber has quit IRC ()
15:57:15 <peter1138> Hmm, coffee, tea, water, or... wine.
15:57:43 <peter1138> I probably need pickpacket's NewGRF :D
16:07:48 *** Wormnest has joined #openttd
16:08:30 <peter1138> I didn't go shopping so I don't have any.
16:40:22 <peter1138> Urgh, usual problem, sprite offsets and bounding boxes are often context dependent, so using generic routines for alignment is unsatisfactory :(
16:43:11 <peter1138> (Some cargo icons are less than 10 pixels high, but the original code doesn't care because it draws it at y 0 relative to the row.
16:43:20 <peter1138> Oh NOW we have beer.
16:47:32 <peter1138> Join the channel :D
16:57:27 <_zephyris> peter1138: I'd totally overlooked them! I'll add them to the scaled GUI to do list.
17:20:27 <peter1138> Well they were never seen outside the station cargo waiting screen.
17:22:10 <locosage> patch packs use them quite more though
18:12:16 *** Tirili has quit IRC (Quit: Leaving)
18:33:19 *** ufo-piloot_ is now known as ufo-piloot
18:38:11 <DorpsGek> - Update: Translations from eints (by translators)
18:42:47 *** Flygon has quit IRC (Read error: Connection reset by peer)
19:34:40 *** Beer has quit IRC (Ping timeout: 480 seconds)
20:40:19 <peter1138> Hmm, template implementation is a pain :(
21:00:59 *** nielsm has quit IRC (Ping timeout: 480 seconds)
21:29:34 <xarick> I also don't understand how they manage more upload than download
21:30:35 <Rubidium> I reckon oversubscription
21:31:51 <peter1138> So that is actually... simpler than the existing filter o_O
21:32:17 <Rubidium> and the fact that things like streaming video use way more of the download than they take of the upload, so if the neighbourhood is streaming a lot then there's less space for your downloads
21:32:31 <peter1138> +46-55... slightly simpler, anyway.
21:34:40 <andythenorth> what did I miss?
21:36:26 <xarick> I should have 120 Mbps/ 12 Mbps
21:36:50 <peter1138> Step 1: Tell them something is wrong. It's probably not intentional.
21:38:46 <peter1138> andythenorth, cheese.
21:51:56 <andythenorth> peter1138: Wensleydale
21:54:42 <_glx_> xarick: at least you have the expected 12 Mbps
22:02:08 *** tokai|noir has joined #openttd
22:02:09 *** ChanServ sets mode: +v tokai|noir
22:08:53 *** tokai has quit IRC (Ping timeout: 480 seconds)
22:10:27 <peter1138> Now it's just over-done.
22:10:34 <peter1138> Information overload.
22:13:28 <talltyler> I dunno, some people are visual people who might want to look at the icon instead of reading the wall of text options
22:26:44 <bungus> peter1138[d]: I'm not a visual person, but I feel icons like that make it easier to find which button to press at a quick glance. I've memorized the cargo sorting listing by now but since they aren't really alphabetical sorting, icons like that could help
22:28:23 <bungus> icons at the front would be better for what I'm talking about though, would be consistent with things like vehicles or trains in the buy list or icons for placing roads/rails
22:30:18 <peter1138> That's a pain because the icons are not fixed width.
22:30:37 <peter1138[d]> Maybe I can template that.
22:34:38 <michi_cc[d]> Icons at the front would probably mean to get the biggest width of all shown icons first, and then maybe center-aligning them.
22:34:41 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
22:36:06 <peter1138> Yes, basically instead of just telling it to use the icon, I'd have to get the maximum width of all items first, then override the width for each item.
22:36:13 <peter1138> Drop down items are very independent.
22:36:25 <bungus> oh god the icons aren't all the same width?
22:36:39 <peter1138> No... you can see that in the screenshot :)
22:37:01 <bungus> yeah, I thought they were just cropped with a lot of whitespace though
22:37:15 <michi_cc[d]> Well, the icons can be a total mix from umpten NewGRFs. And NewGRFs and "standards" is a whole different can of worms 🙂
22:37:54 <bungus> definitely, insert the xkcd about standards here lol
22:38:18 <peter1138> NewGRF authors can't come up with standards.
22:38:24 <peter1138> See the road types thread :D
22:38:35 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:38:52 <peter1138> Anyway, this measuring IS already done for the toolbar menus with icons, so it's not impossible.
22:39:07 <peter1138> But making it right-aligned means I don't need to bother \o/.
22:39:54 <peter1138> Although I have thought of a possible way with this new mixin component system.
22:45:12 <bungus> do you have a link to the source of the C mixins you are working with? Ive seen some written in java before and have a basic understanding of how they work, and I'd like to see what it looks like in C
22:49:46 <peter1138> I'm using CRTP style mixins. "Curiously Recurring Template Pattern". It's C++, not C.
22:52:40 <peter1138> Almost wants to be right-aligned within that space.
22:52:50 <peter1138> OR just make it right-aligned on the right ;)
22:53:06 <talltyler> I vote right-aligned on the right
22:54:24 <peter1138> Yeah, I'm not convinced on the left is better even if it is quite simple.
22:57:34 <bungus> peter1138: thanks 👍 I'll look into it
22:58:01 <bungus> right aligned totally looks better when you compare them side by side
22:58:32 <bungus> all that empty space is begging for something to go there
23:05:12 <peter1138> This is a 'major' inconsistency. All the existing places that show an 'icon' show a coloured rectangle that corresponds to the graph line colour.
23:05:26 <peter1138> Oh, and the map legend.
23:08:47 <peter1138> Reading the second link which distinguishes between CRTP or mixin, I'm using CRTP.
23:08:50 <locosage> iirc inconsistent width is only a problem with original graphics
23:09:04 <locosage> opengfx icons seem to be the same
23:09:14 <peter1138> Original graphics is the only one that matters.
23:09:56 <locosage> very few players actually use it
23:09:59 <peter1138> Yes rEALly. Anything else can be changed.
23:10:36 <peter1138> And NewGRFs can define their own icons with whatever dimensions they like.
23:12:29 <xarick> let me test something. I chatted with Bing, and apparently I can share conversations
23:12:36 <bungus> peter1138: yeah the first one goes over crtp in detail, it's nice to see it compared to a Mixin to see the difference
23:15:30 <peter1138> So this implemention uses virtual methods that are chained through to the base class.
23:16:26 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
23:24:10 <bungus> peter1138: alright I think I understand it, thanks for explaining
23:25:13 <xarick> Why do I have a feeling the links Bing Chat propose are articles also written by AIs?
23:25:55 <bungus> So the CRTP modifies the base class while the mixin leaves it untouched and is more of a patch in thing?
23:33:16 <peter1138> Something like that.
23:33:25 <peter1138> I'm learning this as I go ;)
23:33:28 <xarick> > MB stands for megabytes or megabytes
23:33:28 <xarick> > Speed rate conversion bandwidth: 128KB/s=128×8(Kb/s)=1024Kb/s=1Mb/s
23:33:29 *** xarick has quit IRC (autokilled: Possible spambot. Mail support@oftc.net if you think this is in error. (2023-11-12 23:33:29))
23:34:32 <peter1138> Your IRC presence was just autokilled for being a spambot.
23:50:03 <bungus> yes, this channel is linked to an IRC server and your user in the IRC server got kicked
23:50:40 <bungus> the IRC server makes an IRC user for each discord account to enable sending messages to it
23:52:53 <bungus> heres what it looks like
23:57:36 <bungus> by the way, what bot or system are you guys using to mesh the irc server and discord channel?
23:57:57 <bungus> ive seen a few implementations but none that are this clean
continue to next day ⏵
