Smalltalk with James Robertson

The choices for database back ends to Smalltalk are pretty open - Objectivity just pushed a new release, and it's had Smalltalk bindings for years.

Technorati Tags: objectivity, database, oodb

I've always thought that multiple processes scaled better than multiple threads, and it looks like Chrome has inspired Apple to think the same thing:

WebKit2, which now runs browser elements as separate processes, much like Google Chrome. Actually, Apple's devs say it goes a little farther than Chrome, since the process model is built into the foundation so other non-Safari clients can use it.

Technorati Tags: webkit, apple, safari

I like this post, because it makes a point that a lot of web developers really need to understand: you aren't going to have massive scaling issues:

The sooner your company admits this, the sooner you can get down to some real work. Developing the app for Google-sized scale is a waste of your time, plus, there is no way you will get it right. Absolutely none. It's not that you're not smart enough, it's that you do not have the experience to know what problems you will see at scale.

He's making that point in the context of NoSQL databases (vsersus SQL), but it holds for a lot more. I'm reminded of an anecdote a friend related to me years ago when he just started a new job. He was being shown a web project with a 3 tier architecture, failover capabilities, expensive software all around. When he asked how many users of the system there were going to be, the answer was "3 or so now, maybe 10-20 later". When he suggested that they just use something like Access, they treated him like some kind of heretic, and hustled him along to a different project.

That's the kind of thinking that powers a lot of fake work in this industry...

Technorati Tags: scaling

Way back when data was stored on floppy disks (and then hard drives measured in tens of megabytes), a decision was made on formatting - creating 512k blocks for data storage. Fast forward to today, and that seems ridiculous - but the standard is still there (especially for Windows users), and it's about to cause a bit of grief for people still using XP - hardware vendors are moving to 4k blocks to eliminate some of the wasted space issues:

To help Windows XP cope, advanced format drives will be able to pretend they still use sectors 512 bytes in size. When reading data from a drive this emulation will go unnoticed. However, said Mr Burks, in some situations writing data could hit performance.

Never mind whether you should move from XP to Windows 7; what this really points out to me is just how persistent some early decisions end up being. A mostly arbitrary choice in how data is formatted on a storage medium back in the early days of the PC is still with us. Legacy hierarchical databases are still with us. Choices being made now about how to communicate data over HTTP (like JSON) will still be with us twenty years from now too - because lots and lots of stuff will be written to those APIs.

When you sit back and think about it, it looks like a lot more thought should go into this stuff :)

Technorati Tags: legacy

I like this article by Dare Obasanjo on the relational DB/NoSQL divide - and his summary makes a great point, I think:

For these reasons I expect we'll see more large scale websites decide that instead of treating a SQL database as a denormalized key-value pair store that they would rather use a NoSQL database. However I also suspect that a lot of services who already have a sharded relational database + in-memory cache solution can get a lot of mileage from more judicious usage of in-memory caches before switching. This is especially true given that you still caches in front of your NoSQL databases anyway. There's also the question of whether traditional relational database vendors will add features to address the shortcomings highlighted by the NoSQL movement? Given that the sort of companies adopting NoSQL are doing so because they want to save costs on software, hardware and operations I somehow doubt that there is a lucrative market here for database vendors versus adding more features that the banks, insurance companies and telcos of the world find interesting.

I have to agree there - the needs of the enterprise seem to be diverging from the needs of the large scale website rapidly. Mind you, most websites are not large scale things - for every Digg or Twitter, there are tons of sites that get in the hundreds or low thousands of pageviews a day, and scale just fine with databases or even flat files. I think it remains to be seen whether the NoSQL movement will get much attention beyond the huge scale sites.

You have to recognize what era you live in - and as much as it might be nice to hearken back to some presumed "golden age", it's just not coming back. Where am I going with this? To Mike Taylor's lament about the mundaneness of modern software development:

Today, I mostly paste libraries together. So do you, most likely, if you work in software. Doesn't that seem anticlimactic? We did all those courses on LR grammars and concurrent software and referentially transparent functional languages. We messed about with Prolog, Lisp and APL. We studied invariants and formal preconditions and operating system theory. Now how much of that do we use? A huge part of my job these days seems to be impedence-matching between big opaque chunks of library software that sort of do most of what my program is meant to achieve, but don't quite work right together so I have to, I don't know, translate USMARC records into Dublin Core or something. Is that programming? Really?

I rather suspect that as cars moved from being mostly mechanical to being highly dependent on software the same kind of lament went up amongst car enthusiasts. There's really no going back though; just as I have no real interest in gapping my own spark plugs, I have no real interest in writing my own version of malloc(). Even if I did, outside of being a harmless hobby, who the heck (outside of a handful of OS developers) would pay me to do it?

The world turns, and life moves on. The time for building every little library by hand is gone, and it's not coming back. Personally, I'm happy about that; I just never got a thrill out of low level grunt work :)

Having said that, I can tell you where that level of creation is still going on: game development.

Twitter seems to be moving to Cassandra - a NoSQL database that was largely developed by Facebook. So says Ryan King, a developer at Twitter:

Ryan King, an engineer at Twitter, today told the blog MyNoSQL that the social networking company plans to move from MySQL to the Cassandra database for what he called its resilience, scalability and large community of open-source developers.

This puts the big three social media sites - Facebook, Twitter, and Digg - on Cassandra (or in the process of moving to it). Is this the start of a sea change in data storage, akin to the move from hierarchical to SQL a generation ago?

Technorati Tags: cassandra, nosql, storage

Live by the meme, die by the meme: As a commenter points out, this is a two year old story that I posted on without noticing that rather salient fact. Well, that's what corrections are for - nothing to see here, move along :)

If this TechCrunch story pans out, you can stick a fork in Twitter - they'll be dead soon enough:

We're hearing this from multiple sources: After nearly two years of high profile scaling problems, Twitter is planning to abandon Ruby on Rails as their web framework and start from scratch with PHP or Java (another solution is to stick with the Ruby language and move away from the Rails framework).

The problem is this: swicthing languages/frameworks is a very "enterprisey manager" type decision. Down in the development trenches, it means multiple man years of the following:

• Standing still while the new system is built from the ground up.
• Faults in the existing codebase will go largely unaddressed, as all the effort will be focused elsewhere
• Competitors will have a lot more room to steal users, as it'll be obvious that Twitter is standing still

This isn't a solution; it's a way to circle the drain while others feed off their carcass The worst part is, the management team that made this call (if it has been made - it's rumor now) will walk away clean, blaming "developers" for the failure...

Technorati Tags: RoR, twitter, social media

I've been wary of the Oracle installer tools - especially on Linux - for years, ever since the adventures I had installing on RedHat back in the early 2000's. I would have thought it would be better by now, but I just saw this go by in the Smalltalk IRC channel:

recently I had a problem reading lots of records from PostgreSQL, because the VW driver did not support streaming. So I tried to install Oracle on Ubuntu 9 today.... turned out that implementing streaming was far easier than installing Oracle

That's how I felt about it back then, too...

Technorati Tags: oracle, database, linux

I find it interesting that Oracle feels it's necessary to put out a "Java isn't dying" video. What I find disturbing is the color scheme on that page. Seriously, I felt like I was back in the Deep Roads in Dragon Age Origins...

Technorati Tags: java, sun, oracle