07:59:02 <dan-shearer[m]> #startmeeting 07:59:02 <lumosql-meetbot> dan-shearer[m]: Meeting started at 2022-09-30T07:59+0000 07:59:03 <lumosql-meetbot> dan-shearer[m]: Current chairs: dan-shearer[m] 07:59:04 <lumosql-meetbot> dan-shearer[m]: Useful commands: #action #info #idea #link #topic #motion #vote #close #endmeeting 07:59:05 <lumosql-meetbot> dan-shearer[m]: See also: https://hcoop-meetbot.readthedocs.io/en/stable/ 07:59:06 <lumosql-meetbot> dan-shearer[m]: Participants should now identify themselves with '#here' or with an alias like '#here FirstLast' 07:59:20 <dan-shearer[m]> #here Dan 08:00:17 <gabby_bch> #here Gabby 08:00:19 <dan-shearer[m]> #meetingname Bihebdomadaire Catchup 08:00:19 <lumosql-meetbot> dan-shearer[m]: Meeting name set to: Bihebdomadaire Catchup 08:03:50 <bjk621[m]> #here Björn 08:04:30 <dan-shearer[m]> I'm going to bet that Labhraich is making coffee :) 08:06:24 <rubdos[m]> #here Ruben 08:07:14 <dan-shearer[m]> Morning Gabby 08:09:02 <gabby_bch> Good morning 08:09:14 <Labhraich> #here Claudio 08:09:52 <Labhraich> Dan loses his bet. I was *drinking* coffee 08:10:45 <dan-shearer[m]> ok and Ruben will be here very shortly. 08:10:57 <dan-shearer[m]> looks up previous minutes 08:11:28 <Labhraich> Minutes in the sense of record of meeting, not in the sense of the above Scottish minutes, presumably 08:12:01 <dan-shearer[m]> #topic People and admin 08:12:33 <dan-shearer[m]> #link https://lumosql.org/meetings/Brief_Summary_of_LumoSQL.20220916.0801.html are the last meeting notes. 08:13:00 <dan-shearer[m]> #info Dan will be in Brussels 11-13 October, visiting VUB and other things 08:14:20 <rubdos[m]> We will be welcoming Dan for tea and lunch! :-) 08:16:00 <dan-shearer[m]> valldrac did tell me he would come to the next meeting, which is this one right now, but I realise he isn't on Björn's email list so he would not have known. 08:16:23 <dan-shearer[m]> #action Björn to add valldrac to email list for these meetings 08:16:43 <dan-shearer[m]> #action Björn to remove Tom from email list because that is what Ruben said Tom asked for 08:17:27 <bjk621[m]> #accepted Let's sort the email for valldrac outside 08:18:26 <dan-shearer[m]> #info valldrac and Dan continue to work on aspects of Signal Server (and valldrac on the Molly Signal client), which overlaps with LumoSQL as previously discussed. A first push of code with documentation and analysis is due soon. 08:19:27 <dan-shearer[m]> Any updates from others about who is working where or what? 08:19:53 <rubdos[m]> #info Martina is finishing her thesis due next week 08:20:52 <dan-shearer[m]> Is any external review of her SSS work welcome before it is made public, or does that all happen internally? 08:21:05 <dan-shearer[m]> s/SSS work/thesis 08:21:34 <rubdos[m]> That's out of scope for the thesis 08:21:53 <rubdos[m]> remember: the ABE-SSS serves purely as a proof-of-concept backend 08:22:13 <rubdos[m]> in a real world situation, we'd need a better/more real ABE backend 08:22:23 <dan-shearer[m]> But also remember: Dan wants to be sure he understands the principles at work here 08:22:33 <dan-shearer[m]> Because if he can, anyone can. 08:23:33 <rubdos[m]> Ah I read over your s-expr 08:23:44 <rubdos[m]> It happens internally, but with Martina's permission you can review it too 08:26:24 <dan-shearer[m]> Sounds good. 08:27:15 <dan-shearer[m]> Oh yes and Labhraich noticed that in 2023, Brussels is having a real-world FOSDEM again. Presumably someone has noticed that libre software can't get viruses or something 08:27:55 <Labhraich> I also noticed that hotels in Brussels are already more booked on that weekend than on the previous of next weekend. Don't think it's quite a coincidence 08:29:00 <dan-shearer[m]> #info Dan has got a quote for a group booking for 5 days around FOSDEM, and has passed that to Björn. There may be a way to make that happen. 08:29:18 <rubdos[m]> Another little aside: I'm teaching at 1300, so I'll need to leave 11:30 in order to get lunch before the masses gather. 08:29:31 <dan-shearer[m]> Anyone else is welcome to participate in this discussion, but for now I'm assuming #lumosql people mostly find that very boring. 08:30:17 <Labhraich> #info FOSDEM on 4 and 5 February 2023 for your calendar 08:31:08 <dan-shearer[m]> And valldrac you are included in that if you wish btw 08:36:34 <dan-shearer[m]> Ok that sounds like all the admin sort of work. I have also been looking at how easy it is to include Gothenburg if I am going to be in Brussels anyway btw. Flights and prices seem quite volatile right now, trains less so. 08:37:00 <dan-shearer[m]> Next topic i guess, Ruben De Smet are you ok to chat about PE-SQL-related things? 08:38:55 <rubdos[m]> sure 08:38:57 <rubdos[m]> not a lot to say 08:38:59 <rubdos[m]> but I can 08:39:05 <dan-shearer[m]> #topic PE-SQL 08:39:30 <rubdos[m]> Sorry for the delayed response 08:39:32 <dan-shearer[m]> This topic is about PE-SQL, Ruben's first-born child 08:39:41 <rubdos[m]> SQL-PE, but potato potato 08:39:47 <rubdos[m]> #url https://gitlab.com/etrovub/smartnets/sql-pe 08:39:47 <lumosql-meetbot> rubdos[m]: Unknown command: #url 08:39:50 <rubdos[m]> #link https://gitlab.com/etrovub/smartnets/sql-pe 08:40:42 <rubdos[m]> #info SQL-PE currently implements creating roles and signing roles, and signing grant statements. 08:40:54 <rubdos[m]> That was already the case before last time I was in a meeting. 08:41:22 <rubdos[m]> #info SQL-PE now also has a C interface, `libsqlpe.so`, with a few methods already exposed: https://etrovub.gitlab.io/smartnets/sql-pe/doc/sqlpe/ 08:41:34 <rubdos[m]> #link https://etrovub.gitlab.io/smartnets/sql-pe/doc/sqlpe/ contains the C API documentation 08:41:50 <rubdos[m]> The library and docs are built on CI. 08:42:23 <rubdos[m]> #info Additionally, I've made a Python wrapper around the C interface, in order to easily test whether the C bindings actually work. 08:42:38 <rubdos[m]> None of this already implements encryption, so that would be one of the next logical steps. 08:43:01 <rubdos[m]> Another important aspect is serializing the privileges and roles into something we can store in Labhraich 's hidden tables/columns 08:44:11 <rubdos[m]> I think that's what I wanted to say. 08:44:27 <Labhraich> If they can be represented as a blob, I'll be able to store them 08:44:31 <rubdos[m]> Maybe useful to note: due to Rust name collisions, the C lib is called sqlpe, and the Rust lib is sql_pe 08:44:38 <rubdos[m]> Labhraich: Yep, don't see any problem there. 08:44:44 <rubdos[m]> I just need to do it 08:44:56 <rubdos[m]> and at the same time, make the serialization a bit "standard" 08:45:08 <Labhraich> (In fact, if it can be represented as any sqlite type, but let's stick to "blob" for simplicity) 08:45:11 <rubdos[m]> (and all that without pulling in ugly things like protobufs...) 08:45:35 <rubdos[m]> I might just JSON it out for a first prototype version, ugly as it may sound. 08:45:43 <rubdos[m]> That's pretty easy to do. 08:46:44 <rubdos[m]> There's also a question around whether sqlpe should implement passphrases or should stick to 256-bit keys, but I'll figure out how I feel about that. 08:48:59 <Labhraich> Being lazy, when I need a 256 bit key and want to type a passphrase I just (substr(blake3(passphrase), 0, 32). Possibly adding salt to passphrase first 08:49:01 <dan-shearer[m]> notes for the last time... it's the SQL-PE project, implemented in the sqlpe library. Not "pe" listed first. 08:49:01 <rubdos[m]> ¯\_(ツ)_/¯ 08:49:02 <dan-shearer[m]> I always get that wrong. 08:49:20 <rubdos[m]> Labhraich: yeh, so question is whether we want to avoid that and put it in sqlpe or not 08:49:44 <dan-shearer[m]> What about supporting both? 08:49:47 <rubdos[m]> If we put it in SQL-PE, it can decide what kind of PBKDF to use, if we don't it's a cleaner lib 08:49:54 <rubdos[m]> dan-shearer[m]: I'm leaning towards that, indeed. 08:50:18 <dan-shearer[m]> Can always deprecate one of them in the future, although I can't quite see why really 08:50:29 <Labhraich> On the other hand if we pollute the lib to it, apps using it have a more standard way to convert passphrases to keys, without having to invent and document their own 08:50:37 <Labhraich> So it needs some deciding on the pros and cons 08:50:50 <Labhraich> s/we/you/ actually. I'm just accepting whatever calls you end up making to my API 08:51:20 <rubdos[m]> #accepted Ruben will figure out how to deal with passphrases 08:52:32 <rubdos[m]> I think we can move on 08:56:41 <Labhraich> On an unrelated thing (but still to do with Lumo) I notice that there's been a new release of sqlite3 yesterday (3.39.4 2022-09-29 15:55:41 UTC) so I'll be running more tests after the end of this meeting 08:57:02 <dan-shearer[m]> For those that may be new here... 08:57:14 <dan-shearer[m]> #info Ruben's SQL-PE project is to show how Predicate Encryption can work for SQL, whether SQLite or Postgres or some other database. The current sqlpe design is a prototype that can be called in many circumstances, and is minimalist, suitable for supporting an academic paper. Production sqlpe implementing the same proven concepts would look considerably different. 09:02:02 <dan-shearer[m]> Ok next topic 09:02:08 <dan-shearer[m]> I'm just working out what that should be 09:02:43 <dan-shearer[m]> #topic LumoSQL Project Summary 09:02:58 <dan-shearer[m]> This is the very short version 09:03:33 <dan-shearer[m]> #topic It can feel like LumoSQL is in a bit of a pause, but that is not correct. We are in fact heading towards a release, just not quite how we had expected 09:03:45 <dan-shearer[m]> #undo 09:03:45 <lumosql-meetbot> dan-shearer[m]: Removed event: c442f3af98a84adc818719661cfc47d5@2022-09-30T09:03+0000 09:04:04 <dan-shearer[m]> #info It can feel like LumoSQL is in a bit of a pause, but that is not correct. We are in fact heading towards a release, just not quite how we had expected 09:04:30 <Labhraich> Well, I feel like I'm in a bit of a pause but I've been working on fixing that 09:06:16 <dan-shearer[m]> #info With Ruben and soon Martina working on sqlpe that is definitely happening, and that is the one single core thing that LumoSQL does which is a new application of computer science. So that's not paused. 09:07:56 <rubdos[m]> We've submitted for funding for Martina 09:08:16 <rubdos[m]> I'm asking all of you to cross fingers and toes, with champagne as result for all if we get it. 09:09:32 <dan-shearer[m]> #info With Claudio's architecture for hidden columns and tables all ready to be tested with sqlpe once it gets a little further as discussed above, and since it is already tested in the previous use cases of this code, this is the great Joining Up Of Ideas. Claudio really advanced the art of SQLite with what he has done there. 09:09:57 <Labhraich> is typing with crossed fingers, and it's not that easy 09:12:04 <rubdos[m]> :) 09:12:18 <dan-shearer[m]> #info The concept of Masterplan (I have a better name for this, which I will ask you all about in the chat later) is progressing. We do need more developers, but the way LumoSQL/sqlpe/Lumions/Molly/SignalServer/etc fits together means that sometimes progress has to happen in other components. This is not on pause, the Masterplan concept didn't exist two months ago. 09:14:03 <dan-shearer[m]> #info As said above, we need more developers. We can't release amazing new tools to support human rights to billions of people without a lot more effort than we currently have. That is why I am running around talking to people who employ developers, explaining that we have something new and they should give some of their developers to us. So that is not on pause, in fact, it is fast-forwarding rapidly. 09:18:58 <dan-shearer[m]> #info One of the better University cryptography departments around is highly aware of LumoSQL. The good people at VUB are well-networked, and while these things progress slowly, as the pile of evidence grows there is a reasonable chance that LumoSQL will become useful to curious people. And VUB is very good at being curious. (I should perhaps be saying ETROS not VUB, but I'm not confident of all the fine details of organisation. 09:19:38 <dan-shearer[m]> So there you are. 09:20:02 <dan-shearer[m]> #topic Any other business and extra action items 09:20:28 <bjk621[m]> nop 09:21:14 <rubdos[m]> nothing from me, I think 09:21:14 <dan-shearer[m]> Sure but within that there is X group and Y group. 09:21:15 <rubdos[m]> (ETRO is a department of the engineering facultity (not sciences) at the VUB university) 09:21:38 <Labhraich> nothing from me to add 09:21:41 <dan-shearer[m]> Perhaps the first real-world application for sqlpe is modelling ETRO 09:21:53 <dan-shearer[m]> wonderful. 09:22:29 <dan-shearer[m]> Thanks everyone 09:22:30 <dan-shearer[m]> #endmeeting