Marten’s VA Smalltalk wish list

Marten Feldtmann posted his personal wishlist for VA Smalltalk for 2011. There’s not much to add here, and I spotted a few wishes that I’d regard as high priority as well. So maybe if I help wishing, the list gets shorter around this time next year:

I would like to get/have

* that with 8.03 we get a pretty stable and up to date Seaside implementation with easy packaging management
* to get an up to date version of Glorp
* to get updated database drivers offering newest features of supported database products
* to get support for at least one open database (eh – what about Postgresql ?).
* to get support for LTS versions of Ubuntu. Actually nothing had to be done – at least I have no see major problems – to make VA runnable under 10.10. But the problems with the latest LTS version (10.04.1) are not going to be solved any more.
* I would like to see a good Linux client, which is a pleasure to work with and does not look like a product somewhere located between Englebart and Smalltalk’72.
* a solution to these problems with source code management among different clients with different code pages
* to have more UTF-8 support
* a new management/developer process making it possible to get earlier access to new versions of Glorp and Seaside – and no need for waiting for next releases in these cases
* even for the most important platform (Windows) we have a complicating installation process
* the – even from the community wished – long awaited open bug databases (Frankfurt !)
* enhanced GUI support for Windows and Linux
* support for Macintosh
* support for better IPCs: either with .NET code, Java code or C
* support for better and asynchronous call-in into Smalltalk independent from the GUI thread
* support for multi-processors and/or multi-threading
* support for interfaces within the Smalltalk language or system.
* much brighter activities at conferences and a better support (activity) via Instantiations forum
* a way for project control with external resources (bitmaps), automatically build management
* easier way of building runtime images (headless and headfull)
* some kind if Intellisense when typing code
* a much better kind of GUI, which make source code browsing within the system even more productive
* enhancements in the GUI builder part: do not store things in the repository, but as source code … make it even more powerful
….

Some of Marten’s wishes are on the agenda for 8.0.3 or the release after that (8.5?). So chances are Marten and I will be happy boys next year
I think Instantiations has come a long way in terms of public appearance. And I am optimistic and full of hope at the same time that the hiring of a new staff member for the development team will help make (at least some) dreams come true…

More on NoSQL and Object databases

There is a nice Collection of opinions and discussion about NoSQL databases, object databases and all kinds of “post-relational” database technologies over at ODBMS Industry Watch.

Since CouchDB, MongoDB and other new DBs are showing up weekly, the question whether RDBMS is really the way to go for all scenarios and whether there are easier, cheapier and more effective ways of persisting objects and documents is heating up. Just about one and a half decades after everybody agreed that OODBMS are dead.

Which they are not. Just take a look at Gemstone/S aka GLASS. Seems they are generating quite some buzz. Smalltalk has been a playground for OODBMs for at least two decades, apart from gemstone, there surely are a few more that are still around:  Magma and VOSS come to mind.

There also are some adapters for interfacing Smalltalk to databases like Amazon’s SimpleDB (project Cloudfork) or other RESTful Web Services-capable Databases (CouchDB).

So if you want to combine the highly productive and stable programming language Smalltalk with databases that fit nicely into the object paradigm and let you forget about relational restrictions (which of course also bring quite some benefits), feel warmly welcomed in the Smalltalk community, there are lots of choices for you.

In the above-mentioned article, I had a great laugh about this one:

” it is an interesting discussion. I have heard the term “NoSQL”. I did find the comment about relational databases not supporting key/value stores amusing: “…and index key/value based data, another key characteristic of “NoSQL” technology. The schema seems very simple but may be challenging to implement in a relational database because the value type is arbitrary.”
In Oracle, one simply needs a table as follows:

CREATE TABLE key_value (
the_key NUMBER,
the_value BLOB);

There you go! Key/value. How much simpler can you get? “

That’s clearly a case of “best of both worlds” !!!

Unfortunately, I know a few projects that really use such a DB design.
Unnecessary to say that using such a model throws away most benefits of both techniques… So, please read it as a joke, or at least don’t call me to fix your project, it’s going to be expensive ))