Dyalog Forums

Greetings,

I believe there have been several posts about the "standard" DCF file system tools ⎕FREAD TN,CN

This basic DCF toolset works great for simple single threaded applications.

However, I have been working on some parallel APL session prototypes ... and have a need to truly SHARE an APL Component file system between APL sessions on the same server.

Or, better, leveraging a central APL component file system ENGINE via an API (like a CLOUD based "big brother" to the DCF tools).

I thought I saw a special Dyalog file system product that might be a great solution for this APL database purpose.

I COULD use SQL, and convert the Dyalog APL array structures into character BLOBS and store them using SQL Tables and Rows.

What do you recommend for a robust commercial-quality shared APL-DATA database engine?

Thanks!

//W

Hi Woodly,

the separate product you saw is DFS and you can find out more about it here.

For a quality-recommendation you'd probably have to say a bit more about
- user-numbers
- functionality (i.e. will they all share the same components with r/w-access or are there few writes and many reads etc.) - and certainly somebody with more expertise will come up with even more questions ;-)

:-)

Hi Woody,

I'm probably not understanding exactly what you want to do, but isn't ⎕FHOLD what you need for sharing component files among different APL processes on the same machine? I think that is the purpose of ⎕FHOLD. I think one of the driving ideas behind the DFS system that when you have processes on different machines attempting to share files on a central server, using the ⎕F family of functions does not work so great.

That begin said, I am in the process of open-sourcing my pure dyalog RDBMS, FlipDB. Not sure it qualifies as "a robust commercial-quality shared APL-DATA database engine?" but it might be something you want to look at.

Paul