Fork me on GitHub
#xtdb
<
2019-06-11
>
hoppy22:06:17

so quiet in here. Has anybody started trying to use crux in anger that isn't a juxt'er?

refset23:06:33

@hoppy hey, I haven't looked again in the last 4 weeks, but the last time I searched around on GitHub I found a few interesting things: https://github.com/mjul/crux-db-lab https://github.com/zirmite/crux-fun (just a template, by the looks of it) https://github.com/harold/docbrown ("An application for time-traveling source exploration")

hoppy23:06:11

I have a potential donor project that I'm giving some serious consideration to trying out on crux.

hoppy23:06:17

It's backing store is pure s**t, maintained for compatibilty/interop with an ancient (circa '70s) desktop app. It's sort of a CRM/ERP system, and has probably 100'ish ISAM tables.

hoppy23:06:31

need to move to something, and this is about the neatest thing I've seen in awhile.

refset23:06:56

Wow, sounds intriguing! If there's anything at all we can do to help you figure things out we're very happy to chat 🙂

hoppy23:06:21

I got vexxed using DT for exactly the bi-temporal issues that seem to drive crux. I have historical (log-style) data that has to be fit back in, and I have to do it piecemeal as I go.

hoppy23:06:49

I was pondering how to weave all this together against a TX timeline, and that made my head hurt.

refset23:06:41

🙂 yep, I'm the Offering Managing (a.k.a. product manager)

hoppy23:06:50

oh, star of he podcasts.

hoppy23:06:13

nice job on daniels - got through that this am

refset23:06:06

Thanks! I am super happy with how well-rounded that one turned out

hoppy23:06:53

I'm psyched to get into it. We've been about 3 years dragging this boat-anchor around getting to the point of being able to turn off the old system

refset23:06:13

What you described about CRM/ERP migration reminds me of some great comments I read on that "Salesforce is buying Tableau for $15.7B" HN thread.

hoppy23:06:30

I'm a niche in a niche in a niche

refset23:06:59

I think we all are! I have to run now, but I looking forward to chatting again soon, and feel free to ping me if you fancy a video call sometime

hoppy23:06:23

so despite all the PR, I'm still unclear on one point. If I have, say a 20K document, and change one field, what does crux store, the whole thing again, or a delta?

hoppy23:06:32

run if you must, I was mid-sentance on that.

refset23:06:40

Haha, that's an easy one

refset23:06:45

It stores the whole thing again

hoppy23:06:55

ok, that gives me planning fodder.

refset23:06:06

Maybe you can run Kafka on ZFS 😉

hoppy23:06:16

don't make big docs, make little ones that relate

refset23:06:19

yep it's a balancing act for sure, but we're hoping to create a subgraph decorator that will make splitting your documents apart (and accessing them again) more seamless

hoppy23:06:43

My stuff is currently fragged to the max, so it would likely fit the mold

hoppy23:06:00

It's from a world where 10MB was a big deal

hoppy23:06:59

Go run away since you need to. We need to talk more about what I should do for a backing store. I managed to get rocks workiing on linux, but I need a windows story.

hoppy23:06:25

Kafaka too much for this. We are 5-30 user, single server stuff.