4 ms·

Is this really an important feature for DB users here? I built my own hobby database system that was column oriented (using the metadata tagging system I invent

Is this really an important feature for DB users here? I built my own hobby database system that was column oriented (using the metadata tagging system I invented for a file system replacement I built). It does lightning fast analytics and even outperforms SQLite in a bunch of other operations. I posted a benchmark video on my YouTube channel and made a free beta download available; but the response has been lukewarm at best. https://YouTube.com/@didgetsdemos https://YouTube.com/@didgetsdemos https://www.didgets.com https://www.didgets.com

lukevp 3y ago
It’s not open source (or even source available), and it’s targeted at users who, what, use csv files but can’t use excel very well, but are searching for a new revolutionary file system? I don’t really see a use case for your product based on the pitch page, just a lot of marketing speak that (to an engineer) seems like an apples and oranges comparison (eg. You’re running a local db, could be in memory even, and benchmarking it against Postgres, which is an incredibly flexible and comprehensive piece of software with multitenancy, transactions, replication, schema management, query planning, and a full on webserver? How is that the same product? Why not benchmark this against Excel? Cuz I’ve already got excel installed and I can open a CSV and grep for a string and count the matches in like 5 seconds. That seems to be what you’re competing against. If you open source the code and make it a general purpose library (and not windows only) and market it toward developers, you might get some more interest.
- didgetmaster 3y ago
  I never claimed that my project in its current form is a drop-in replacement for Postgres or even SQLite. It still lacks many features that those databases have accumulated over decades of development by large communities. It is not just an in-memory DB. All its data is persisted on disk and I made sure whenever I did a benchmark on any query; that I was doing an apples-to-apples comparison. Even though the beta consists of a Windows executable; there is nothing Windows specific about it. It is built to be cross-platform and it has been tested on Linux. With my small team, we just maintain a single platform at the moment for simplicity sake. We have not yet open-sourced the code as you pointed out; but we have considered it. But if everyone just wants to dismiss it out of hand because it fails to check some box and just focuses on what it can't do, instead of what it does well; then there might not be any point in doing so.

lukevp 3y ago

It’s not open source (or even source available), and it’s targeted at users who, what, use csv files but can’t use excel very well, but are searching for a new revolutionary file system? I don’t really see a use case for your product based on the pitch page, just a lot of marketing speak that (to an engineer) seems like an apples and oranges comparison (eg. You’re running a local db, could be in memory even, and benchmarking it against Postgres, which is an incredibly flexible and comprehensive piece of software with multitenancy, transactions, replication, schema management, query planning, and a full on webserver? How is that the same product? Why not benchmark this against Excel? Cuz I’ve already got excel installed and I can open a CSV and grep for a string and count the matches in like 5 seconds. That seems to be what you’re competing against. If you open source the code and make it a general purpose library (and not windows only) and market it toward developers, you might get some more interest.

didgetmaster 3y ago
I never claimed that my project in its current form is a drop-in replacement for Postgres or even SQLite. It still lacks many features that those databases have accumulated over decades of development by large communities. It is not just an in-memory DB. All its data is persisted on disk and I made sure whenever I did a benchmark on any query; that I was doing an apples-to-apples comparison. Even though the beta consists of a Windows executable; there is nothing Windows specific about it. It is built to be cross-platform and it has been tested on Linux. With my small team, we just maintain a single platform at the moment for simplicity sake. We have not yet open-sourced the code as you pointed out; but we have considered it. But if everyone just wants to dismiss it out of hand because it fails to check some box and just focuses on what it can't do, instead of what it does well; then there might not be any point in doing so.