If you are in Håkan's @strangeloop_stl talk right now (or even if you're not), you can follow along with his slides and read all his references:


"Deletion considered harmful. ... People do not delete. The deletion concept was invented to reuse expensive computer storage."

Storage isn't expensive any more... so why do we still delete?

Håkan's talk is up in 10 minutes in Grand F. @strangeloop_stl

Speaking of putting theory into practice (and dialling down from Deep Learning to Light and Adaptive Indexes), we're excited to watch Håkan's talk in Grand F this afternoon at 15:30. See you there!

Deep Learning bug discovery and debugging with Foutse Khomh (Polytechnique Montréal) and Tianyi Zhang (Purdue University).

`DeepSeer` transforms DNNs "back" into FSMs that humans can reason about. Tools to enable Responsible AI — definitely a much-needed field of research.

Very punchy and entertaining talk by @d_chrisbrown2 on the research around the psychology of encouraging developers to use better tools. It feels like a lot of the anecdotal evidence from the past 20 years of ~XP, backed by real data. (+bots) Neat research!

An interesting approach to consistent API usage by Prof. @sarahnadi from the University of Alberta.

Did you miss @refset's workshop yesterday? Never fear! You can still work through some of the exercises he created. Have fun!

RT @juxtpro
We’re in the New York conf room at @strangeloop_stl to listen to @refset on ‘Bitemporal Data Management with @xtdb_com’. Join the workshop to find out why you need .

Tomorrow (today?) at 12:30 Central, @refset will give his workshop at @strangeloop_stl

We'll publish some of the workshop material afterward on xtdb.com/core2/ — stay tuned.


RT @KeithWHare
In the standards world, we have completed a new version of the SQL standard with Property Graph support and next year will complete new property graph database language standard called GQL. twitter.com/pdrmnvd/status/157

It's about 24 hours until a few folks from the XTDB team land in St. Louis for . See you there!


Håkan and Matt are not on Twitter but we swear they do exist.

If you had an Immutable Database, would you bother with steps 2 - 5 at all? Or would you simply let the schema evolve alongside your data?

(Both are valid. Just a thought exercise.)
RT @mscccc
How high traffic sites rename a db column without downtime.

1. Add new column
2. Double write to old/new column
3. Backfill new column
4. Switch reads/writes 100% to new column
5. Delete the old column

Each step is a code deploy

RT @PrinceVogel
Get in the car, we're exploring databases and finding friends

RT @refset
It's been a decade since Craig Baumunk (RIP) outlined "Bitemp 2.0" ...but the revolution is still ahead of us! youtube.com/watch?v=R7kQO5Z2JK

I'm really hyped to be talking about these ideas at @strangeloop_stl this week in the context of our next-gen DB engine: xtdb.com/core2

Upgraded to XTDB 1.22 in production yesterday?


It's Friday. Want to grab a coffee and try the experimental XTDB Core2 pre-alpha on your dev machine today?


XTDB 1.22.0 is ready 🎉 github.com/xtdb/xtdb/releases/

M1 users should have a better time using XT now, and the ingestion speedup improvements will make re-indexing existing databases noticeably faster. Enjoy!

"Setup Emacs as an SQL Database Client"


We had a very compelling demo of this today from @jarohen while discussing SQL UX. We encourage you to try it out!

"... (1) as if all of it were stored at the local node--location transparency--and (2) as if no data were replicated anywhere--replication transparency."

-- E. F. Codd

Relational Database: A Practical Foundation for Productivity, 1981



Show thread

"In the longer term we may expect support for relational databases distributed over a communications network [25, 30, 32] and managed in such a way that application programs and interactive users can manipulate the data:" [3/4]

Show thread

"Such data may include images, text, and miscellaneous facts. An affirmative answer is expected, and some research is in progress on this subject, but more is needed." [2/4]

Show thread

"As it stands today, relational database is best suited to data with a rather regular or homogeneous structure. Can we retain the advantages of the relational approach while handling heterogeneous data also?" [1/4]

Show older

The original server operated by the Mastodon gGmbH non-profit