Rubénerd Blog :)

Ever since I postulated back in April as to what a combined Oracle-Sun company would do, in the back of my mind I’ve also been thinking what it would look like. Now we get a glimpse in the form of the Oracle Exadata V2 server, the second advertisement after their triumphant "Oracle Buys Sun" faux billboard graphic.

Read this post >

It seems as with many technologies, regardless of your own abilities or competency if you’re using the "wrong" implementation you’re perceived as an amateur, and vice versa.

Read this post >

Update: It’s been brought to my attention that not defining primary keys as not null does have an impact after all, refer to Dale McGowan’s comment below. Suffice to say, I will continue to define them as not null from now on!

Much like a musician who despite getting the right sound realises they’ve been playing their instrument wrong throughout their career, I often find I’m not using computer software properly or I’ve been wasting effort. I guess I get myself into a predictable routine that I’m used to.

In this case I found out something about Oracle databases. I’ve been using them for years, but it took a practical class at university yesterday to realise I’ve been wasting time with a simple line of code that was completely unnecessary.

If you do any sort of work with databases you’d probably want to smack me on the head for this, but I didn’t know that primary keys are automatically assigned as not null; presumably because it doesn’t make logical sense and it would break a table to have a primary key that has null values!

Up until Friday this is the schema I would have created and entered into an Oracle database:

create table Jackarse (
  refID char(6) not null,
  name varchar(32),
  grilledCheeseSandwich boolean,
  primary key (refID)
);

And here’s what I’ve since learned I can do instead:

create table Jackarse (
  refID char(6),
  name varchar(32),
  grilledCheeseSandwich boolean,
  primary key (refID)
);

That’s right, both these schemas will result in the same table being created in Oracle. I guess it wasn’t harming anything having the not null command there, it was just unnecessary.

That wasn’t quite the "time capsule" I had in mind…

Having spent the better part of the afternoon and evening here working on fixing broken links caused by my recent abrupt domain and webhost move, I encountered a potential problem to do with posterity. All the world’s problems pale in comparison to it you must understand.

The problem was, internal links between posts on my blog here that were created before the move pointed to addresses at the old URL. With some SELECT REPLACE requests to my MySQL tables here I was able to change all these links from over five years of posts to point to this new domain (well actually this domain is older than the one I moved from, but that’s beside the point).

Once I’d cleaned up these links though, I started getting carried away. I started converting some categories to tags that these older posts had assigned to them, then I started optimising some of the images they contained and reuploading them. Previously I had also added header images and text explaining the context in which old posts were written, and I replaced one entire post from late 2005 with a grilled cheese sandwich I had cooked but dropped on the floor and was subsequently reluctant to eat.

Screenshot of an old blog post from 2004 modified with a header thingy and whatnot

So here’s what I got to thinking: Is it really right to be doing this? I’ve been reading a lot about the concept of a digital dark age (Wikipedia) where original material gets lost online because it’s in a constant state of flux and some material inevitably goes offline never to be seen again.

With this in mind, should blog posts be treated as time capsules? As in, when you create a blog post and publish it, to be honest should they never be changed? In my case I didn’t change any of the content per se, I just changed the links and optimised the images, but the text and the images themselves are the same. I know plenty of my earlier posts have spelling mistakes and the images aren’t aligned correctly, but I haven’t felt right changing them. As far as I care, posts from 2005 were written in 2005 and look the same as they did in 2005.

It’s funny how ideas and quandaries that seem really important when they’re floating around in your head seem extremely important, but when you explain them by writing them out on a blog post they seem so utterly ridiculous in their irrelevance.

Hey, at least I spelt "quandaries" correctly.

Showing my support running OpenSolaris in VirtualBox on my Mac :-)

Well it’s official, Sun Microsystems found a suitor not in IBM, Apple or Microsoft, but with Oracle. I must say on the whole I’m cautiously relieved, though there are some things that worry me. What an ambiguous sentence. Well you know what they say, certainty is hobgoblin of the fool… right? Wait, that’s not how it goes.

Firstly, Java and Solaris absolutely make sense for Oracle; I believe I read somewhere that (paraphrasing) Larry Ellison claimed Solaris was the most effective Unix system out there for use with Oracle infrastructure. I haven’t played with OpenSolaris (or Linux for that matter) to the extent I’ve used FreeBSD but I was really impressed by it’s completeness and quality. I’m also interested in licencing; as a BSD guy the CDDL that has prevented Linux users from adopting ZFS and Dtrace hasn’t bothered me but Oracle’s work with Linux may see it change to the GPL… maybe.

What I’m also interested in is Sun’s free and open source acquisitions and assets, namely NetBeans, VirtualBox, OpenOffice.org and MySQL. If they keep them and continue to support them I don’t see any problem, and one could argue that said projects would benefit from being part of a larger company in the same vein as Linux at IBM. Some brainstorming on each:

• NetBeans would compliment their Java acquisition, but Oracle is a signed member of the Eclipse foundation. Will Oracle attempt to merge it, ditch it or continue to develop it and instead ditch Eclipse?

• VirtualBox at Oracle is a fascinating combination. Could Oracle somehow leverage VirtualBox’s virtualisation technology to more efficently deploy servers with Oracle’s database? They could do it with Solaris too!

• OpenOffice.org is a real wildcard. One could potentially see Oracle leverage OOo against Microsoft, or perhaps they could pull a Novell and instead fork it into an independent stream and an Oracle branded product that could use groupware Oracle develops. The possibilities are extremely intriguing.

• MySQL for me is a major concern. Oracle’s databases and MySQL are light years apart in features and from what I’ve heard scalability, but the gap is much narrower than it was even a few years ago. Is it conceivable Oracle would intentionally cripple MySQL or keep it with only a subset of features so as to not cannibalise their bread and butter? If they do, might we see MySQL forks or even a general move to PostgreSQL?

I’ve done work for Oracle in the past but I still don’t know the inner workings of the company or the details regarding merged product lines (ala PeopleSoft and Siebel). Whatever the outcome, I’ll be watching closely.

As I’ve said here before I have huge respect for Sun and have always wanted a pimped out Sun Workstation to really get stuck into Solaris. In fact if I were offered a Mac Pro or a similarly high end Sun Workstation I’d take the latter just because I’ve used them at university but ave never had one of my own before!

The biggest question I still have though is: what will happen to one of my top technology idols Jonathan Schwartz? Will he get a position in Oracle? Will he continue to blog from there? Or perhaps a more pertinent question would be: will he be allowed to blog from there?

I wish Sun the best in this transition period.

The barely listenable web development episode!

Hashing out and talking through some ideas about the future of Rubenerd.com and how I want to move my five separate sites over to it. Might be listenable for a pleasant background distraction if you mess around with weblog engines, databases and whatnot, otherwise don’t feel bad giving this one a miss. Looking at Dave Wares’ Photo Gallery as an example of where I want to be.

To make it somewhat more interesting, I’ve included some retro audio easter eggs :-).

Download MP3 to listen ↓ 01:12:11 33.1MiB

You can also stream this episode and view its Internet Archive page.