[01:45:15] Yay! We reached our audience! (re @lucaswerkmeister: *hammers “test suite for evaluation engines” subscribe button*) [05:09:32] would it be possible to have subtitles for the presentations? [08:58:33] will JS still be one of the two initial programming languages supported? [13:03:36] Hello. I wonder if there's been a study for existing parallel (commercial) services already offering support for hosting (lambda-)functions, notably Amazon's Lambda, Microsoft's Azure Functions, IBM's Labmda, all termed to be "serverless", i.e. that can be deployed in seconds without creating, configuring and allocating any server or container and sharing a common infrastructure. As well some parallel with e [13:04:14] Hello. I wonder if there's been a study for existing parallel (commercial) services already offering support for hosting (lambda-)functions, notably Amazon's Lambda, Microsoft's Azure Functions, IBM's Lambda, all termed to be "serverless", i.e. that can be deployed in seconds without creating, configuring and allocating any server or container and sharing a common infrastructure. [13:04:14] As well some parallel with existing programming languages, notably functional ones like Haskell or Racket. [13:04:16] And existing code repositories (GitHub, CSPAN, npm...) and why Wikifunctions will be distinct: open licencing, free for use by everyone, serverless but also decentralized (possibly using distributed computing and not necessarily centralized, meaning it would be scalable as well) [13:05:05] Hello. I wonder if there's been a study for existing parallel (commercial) services already offering support for hosting (lambda-)functions, notably Amazon's Lambda, Microsoft's Azure Functions, IBM's Lambda, all termed to be "serverless", i.e. that can be deployed in seconds without creating, configuring and allocating any server or container and sharing a common infrastructure. [13:05:05] As well some parallel with existing programming languages, notably functional ones like Haskell or Racket. [13:05:07] And existing code repositories (GitHub, CSPAN, npm...) and why Wikifunctions will be distinct: open licencing, free for use by everyone, serverless but also decentralized (possibly using distributed computing and not necessarily centralized, meaning it would be scalable as well), and multilingual since the start (unlike all other repositories) [15:16:53] Yes, JS is still the main contender for one of the initial languages (re @lucaswerkmeister: will JS still be one of the two initial programming languages supported?) [15:18:09] alright, thanks! [15:20:30] I’m happy as long as at least one of the languages has a decent GraalVM implementation ^^ (i.e. JS, R, Ruby, Python, maybe WASM would be great candidates from my point of view) [15:40:45] The paper on Arxiv discusses related work a bit, but as your question demonstrates - how does it relate to Haskell, or Amazon Lambda, or GitHub - this is a little bit one of a kind, inspiried by all of these and yet different than any of these. I think if, e.g. someone where to write such an unbiased study, it would make most sense when Wikifunctions is actually launched, otherwise we're just comparing vapo [15:40:46] As well some parallel with existing programming languages, notably functional ones like Haskell or Racket. [15:40:47] And existing code repositories (GitHub, CSPAN, npm...) and why Wikifunctions will be distinct: open licencing, free for use by everyone, serverless but also decentralized (possibly using distributed computing and not necessarily centralized, meaning it would be scalable as well), and multilingual since the start (unlike all other repositories)) [15:42:07] Thanks, I'm noting this down. And yes, I'm almost certain that one of the initial languages, if not both, will be on the list you just mentioned (re @lucaswerkmeister: I’m happy as long as at least one of the languages has a decent GraalVM implementation ^^ (i.e. JS, R, Ruby, Python, maybe WASM would be great candidates from my point of view)) [15:46:22] I don't know. I guess if I made a video myself (now that I know how) and upload it to Commons, we would have the infrastructure to add and edit subtitles? I'm not so knowledgeable about videos and subtitles (re @Nikki: would it be possible to have subtitles for the presentations?) [16:08:16] Back at Christmastime I started [[c:User:Mahir256/subtitle-editor.js]]; it's still buggy in a lot of areas, but it's a start as far as editing goes [16:13:06] Thanks! I guess I have to make the video first [16:13:17] I created some subtitles through aegisub and uploaded to commons [16:13:47] Oh both options are good [16:14:24] here's a sample https://commons.wikimedia.org/wiki/TimedText:Processo_alla_reta_%E2%80%93_Teatro_delle_formiche_teaser.mpg.it.srt [16:15:35] basically I downloaded the video file from commons, loaded in aegisub (which shows both video and audio waveform, then I timed the phrases from the audio, and did the transcript [16:15:56] exported as srt (which is a simple text file) and loaded in commons [16:26:39] I don't have much experience with adding/editing them, I'm usually the one using them, but from what I've gathered, a common workflow is to start by uploading to youtube and copy the text from the automatically generated subtitles to get an initial transcript, then go through and correct all the mistakes, upload the cleaned up transcript to youtube and let it do the timing, which appears to work fairly well, except it's really n [16:28:56] the c3subtitles group uses amara for the timing and has a mixture of transcripts from youtube and trint (but I just looked at trint and whoa expensive, the cheapest plan is €44 a month!?) [17:37:47] Daniel Garijo shared this RDF dump model of Wikidata in Wikifunctions mailing list. It will be interesting to have this figure in Commons. [17:37:58] https://tools-static.wmflabs.org/bridgebot/d33d7498/Screenshot_from_2021_02_19_17_04_12.png [17:38:29] This will allow people with advanced knowledge of SPARQL to query Wikidata information. [17:40:08] The figure that is commonly used (https://commons.wikimedia.org/wiki/File:Wikidata_RDF.svg) is absolutely difficult to read although we reproduced it for our research paper.