Wednesday, February 23, 2005

Projectile Management

The guys at Surity, who are running our week long PM workshop have some excellent skills and resources when it comes to project mangement.
(don't let their appalling website fool you. It's true - the blink tag lives!)

I've learned a whole bunch of really valuable stuff, and it's made me re-consider what I think the really important elements in project management are. So, I'm going to blat them out here.


Build a Great Team
I still think that there is more to be gained from optimizing the team rather than the process. That idea intrigues me - maybe because you can't fill in a template for a great team...most of the focus at this workshop has been around contractors and software deployment management - where you don't get to pick the team, and you have to deal with whoever the customer sends. In an ISV software project , you can choose the team that you want to run a project. It sounds totally obvious, but If you want great software, this is the single most important project element there is. A great team can transcend any process flaws.

Bad news is always best early
Unlike a fine wine, bad news gets worse with age. People like everything to be on track, and they don't like to admit that they might have screwed up. So when things go bad, people often tend to solve the problem by pretending that everything is fine. This is the worst possible outcome. If things are bad, any team member should be yelling and carrying on as soon as they know it. And on a good team, they're just as dissapointed as anyone else, but they know that their team won't persecute them for it.

Slipping is Not Bad
Slips happen - schedules are always future gazing things, and humans are only good at predicting the future in the short term. For instance, I bet you can pretty accurately predict the next three minutes with about a hundred percent accuracy. But once you try to predict a week, it get's fuzzy, and if I ask you to predict the next six months, you'll have no chance at all.

So, a slip is when something you didn't know, becomes known. While this is dissapointing because the project is under more pressure, it's also good, becuase we know that stuff now. We can get a more accurate idea of what this evil task looks like, and ammend accordingly. Like Jim McCarthy says - a slip should be a net positive.

Define, Define Define
Really, this is all project management is. Keep defining everything you can get your hands on, and get those definitons into the teams collective brain. Define Risks, Define Responsibilities,Roles, Tasks, Issues,Benefits, Functionality,Milestones, Budgets, etc. Nearly all the scary problems in project management come from the areas that haven't been properly defined. And you can't just stop defining once the design is done - you have to keep defining everything you can get your hands on right up until ship day. Really, it's a very definey job.

Tuesday, February 22, 2005

death by blog posts

Stilly is having trouble keeping up with his blog feeds.

I was having the same problem, but now I've got it down to under half an hour. (Around 300 posts a day)

My suggestion - use onfolio and create a newspaper with all your feeds, skim through them and click the reading tray icon for stuff that looks vaguely interesting, and mark the rest as read. Then browse the reading tray at your leisure.

Oh, and if you want to read some real death by blog posts, subscribe to the blog of death. Unfortunately, it tends to be somewhat depressing...

Monday, February 21, 2005

How many project managers does it take to change a light bulb

Spent the day in a project mangement workshop with the aim of refining and improving our PM methodologies.

I've never been in a room with so many project mangers before - Admittedly I was the only one with a software development background - the rest of the guys are all your more standard IT deployment/change mangement folk.

Some idle thoughts:

Maybe this PM stuff is one great big marketing beanie that you pull over the head of prospective customers and yell "Look How Professional I am!"

Project Managers by their very nature tend to be control freaks. Just a few weeks ago I was making a joke about old school managment, where you make a plan and nobody believes it except for you, and then everybody lies to you so you can tick boxes on a schedule that became irrelevant as soon as you wrote it. I think most waterfall development methodologies suffer from this problem. But, the agile methodologies suffer from a fuzzy level of schedule that means you can't plan very far ahead. Surely there's got to be some common ground?

There are huge amounts of money spent on this stuff, which often doesn't result in success. Most big IT companies have a brain-mooshingly enormous project methodology that sits in a huge box gathering dust in every project office. And it's main purpose is to fill in tender responses with bogglingly complex graphics. I know that when I worked for the Australian Government, we had a big box of methodology that I used to keep my coffee machine on.

It was very useful for holding coffee machines...

I guess the point is that if you've developed a methodology that has a heavily beaucratic undertone,to the point of giving stakeholders the shits, you're not going to get much in terms of performance and risk management, which is the whole point of said methodology.

Unless it's the beanie thing.

Friday, February 18, 2005

Emperor Gee 5

The friendly folk over at Apple developer relations lent us a G5 for our project. (Our code looks like it will run on pretty much any platform, and now we know it will run on OSX...)

Really, the amount of developer drool in my cubicle is obscene. And it's not at all unfounded. It sure is a fancy machine. And the 20" cinema display is simply the best one I've ever seen. If I had 7,000 spare dollars I'd buy one...

At work we use TRIM for tracking bugs - It's super extensible, and has all the features available in any bug tracker(plus a gazillion more) . But, the amount of legacy theory and added complexity made me investigate what other alternatives are available. BugTracker.NET is a great, open source variant that looks to me to be heavily inspired by FogBugz. If you need a quick free windows bug tracker - or even if you're considering FogBugz - you should check it out.

Given that Joel Spolsky seems to be building some kind of developer nirvana over there at Fog Creek, It makes me wonder when his company will actually do something...Surely you can't grow a hugely successful software company armed with bug tracking and dinky content management? On that note, I was speaking to a developer of a popular developer tool about SourceGear's Vault product - sadly, he had nothing nice to say about working with or integrating with Vault.

Eric Sink is another one of these great and talented bloggers (and CEO of SourceGear), and a guy who really really gets the blogging as marketing thing. Those guys are both doing an awesome job of building developer loyalty by appearing as pragmatic leaders of the development community. I' m sure If I posted this stuff on the JOS forum, there'd be no end of developers (who have nothing to gain by defending them) flaming me because they've bought mindshare from those guys..

I love that story about the emperor's new clothes. Not to say that their products aren't good - just that they mightn't be as amazing as the hype suggests.

More Marketing and Policy...

Tuesday, February 15, 2005

Look! Up in the sky...

Had to post this picture of me and my brother clowning around on the headland:

Ruby's caption was better:

"Gordon and Graeme Practice for the K-Mart Catalogue"

Unstructured Thoughts on Unstructured Search

If you were wondering what's on TV in California, check out Google video. Apparently they're indexing the supetext captions that are transmitted in the black bits of our TV channels. Man - those guys are developing a finely tuned indexing radar. If there's content anywhere, (even flying through the atmosphere!) -seems they'll find a way to index it.

This makes me speculate on the future of companies that have profited for years by providing structured storage (FileNET,Documentum, OpenText, TOWER Software) etc.

I mean, when you can triumphantly return with exactly what you wanted from a big pile of random stuff, do you really need to structure it properly? Let's use the following half-baked analogy:

My study is full of vaguely important pieces of paper. Becuase deep down I hate all of them, I tend to treat them fairly carelessly. This means I generally just chuck them through the study door in order to minimize the amount of time I have to spend thinking about them. Sadly, every now and again I have to retrieve one for some random bureaucratic purpose. This usually involves swearing and yelling, and lots of looking at papers that I don't want.

If I could use these two approaches(structured vs unstructured storage) on my study, one would shake a bony librarian finger at me and say "You should've filed all your papers in folders, and categorized them appropriately. I have no idea where you car insurance policy is. " After a bunch of searching, I'd eventually find it - probably with lots of swearing and yelling.

Whereas the unstructured index mad loons at Google would just hand it to me. (along with a bunch of other unrelated cat insurance policies and so on. )Minimal swearing and no chastising.

So - is unstructured storage the way of the future? If that means less filing, I'm all for it. Now, If I could only get a googlebot to index my study...

Chicken Restoration Program

In an effort to replace our previously massacred chickens, I built a new chook house on the weekend. I bought a new angle grinder for the princely sum of 19 dollars, and got a little carried away...

Seriously, who would make an angle grinder for 19 dollars? (If I made one, I'd charge you a fortune for it...) anyway, lots of loud sparky fun. My god you get dirty when you brush five years of crusted on chook poo into a fine dust haze... I thought I had a great tan until I started to sweat.. and my nose cleared....

As yet no word from Stilly -I'm afeared that Microsoft may indeed have lured him to his doom. Simon has some interesting Newspaper Clippings that may shed some light on his blogging silence.

Tuesday, February 08, 2005

Penguins, Mouses and Tomato Sauce


My mouse is covered in tomato sauce. Wireless mouses are a bad, bad thing. If this mouse had a cable, nobody would've put it on a plate of tomato sauce. It does explain it's slow, gooey movement however. That's one problem solved.

Rory wants to visit. I wonder if I can convince TOWER to bring him out for our next conference? A lunatic in a bunny suit can only be good for business.

My good friend Stilly (Stillbert as he's affectionately known) is off to Seattle to see if he can get "a job" at Microsoft. (My other friend, Simon thinks that it's really part of a grand plan to collect all the linux hackers and detonate them. ) For some strange reason I had this vision of a giant penguin dancing down Microsoft Way playing a pan pipe, with a crowd of pasty linux nerds in a conga line bhind him...

"Yeah - we saw this cool Tux , and like we followed him coz he was dancing down the road in front of the campus at Microsoft."

And then Tux opened a door in the side of Building 19, and all the linux hackers disappeared were never seen again...Bwahaha!

I hope nothing like that happens to him. But man, it would be ever hillarious if it did.

Why is there no implementation of Secure Sockets in the .NET framework? That is retarded. (Of course, it will be fixed in Whidbey. Many things, including world hunger will also be fixed by Whidbey. )

Friday, February 04, 2005

Toy Lust

I really really really want one of these.

And isn't that just the greatest bit of flash marketing you've seen for a while?

Or am I just such a nintendo fanboy that I think it's all good regardless...

Thursday, February 03, 2005

Policy and Marketing

Thinking about this previous post led me to question how easy it is to fool people. Following the architecture astronaut mentality this led me to abstract all human behaviour thus:

Seems to my potty head that everything that people work on can be summed up in these two things - Policy and Marketing.

Policy is what people do, what they will do and so on - Marketing is how they will percieve what gets done. Now, we all know that perception is easier to manipulate than reality - it's much easier for me to convince you that I had a wee on the top of Mount Kocsziusco than to actually do it.

Add to this the generalisation that all people are fundamentally lazy. So they often tend to work more on perception than reality. Which is why it's occasionally easy to fool people.

In my line of work, I work with a bunch of engineers. I like engineers. One of the things I like about them is that they don't do bullshit. They only care about what really gets done. (Often this makes engineering types very difficult to talk to, becuase they don't see the point in talking when there is nothing to say. They do, of course have a good point.)

This endearing quality is the same one that makes them wear shorts, socks and sandals.

Initially, they tend to be easily influenced by marketing and perception because they assume that what people are saying is true. After a few bad experiences, they end up with a finely honed bullshit detector, which leads to the cynicism so popular these days....

On the other side of the fence, there are people who don't do anything - just manipulate the way people see things that have been done. Look at advertisers, marketers, press secretaries, PR companies and Journalists, Lawyers and Accountants (well, dodgy ones). What actually happens is pretty much irrelevant. These guys spend their lives manipulating the way that people percieve things. And, they generally make a shitload of money - becuase their efforts to change people's minds are effective. The reason people pay for advertising is because it works.

If you look at all public institutions, whether they are companies, governments or churches, political parties or freaky cults, they all have two clear divisions - people who do stuff, and people who persuade people the stuff is good. (In software land, this is clearly illustrated through Eric Sink's brilliant closing the gap articles)

My Friend Matt is a Roman Historian. He's also a D&D role-playing, web developer and skilled IT professional (read:card carrying nerd) who's never heard of slashdot, but that's a different story.

Anyway, when I saw him last week he gave me a roman coin from about 300 A.D. I've always liked this coin, becuase the reverse inscription has a Roman soldier stabbing a Persian, and the inscription around the outside reads "Happy Days Are Here Again!" (except in latin)

Matt tells me that in actual fact, the Romans were being pretty badly beaten by the Persians, but that didn't stop them from actively trying to persuade people otherwise. Nothing new here.

One side of the coin is policy, one side is marketing...

(oh, and if you're looking for amazingly great roman coins - Trust, me, they're really, really good.)

Tuesday, February 01, 2005


Miles Davis had an album called MilesTones. I always thought that was pretty cool.

We finally hit our first major milestone at work yesterday - the M1 milestone was based around getting an established codebase and tools completed - test and communications plans and a production environment along with processes for promotion.

Here's me thinking that the development methodology we're using is all fancy pants and unique, (and that maybe one day I could write a book about it and retire to Barbados), when some irritating JoS Poster mentioned something called SCRUM. SCRUM is pretty much the development methodology we're following on Project Tremble. We've got all the good bits of XP, the good bits from MSF, and none of the silliness. We're having daily 10 minute meetings instead of a weekly breifing. We're pretty much answering the three SCRUM questions daily (although not so formally).

The thing that really doesn't sit quite well with me about these agile methodologies, is that they are so hard to schedule. You just bite off a chunk, get cracking, and there's an overwhelming (well, maybe just whelming) sense of "It'll be done when it's finished." Given that you always seem to end up with a fixed ship date, when it comes time to ship, right now I'm not entirely sure exactly what will be ready...

Ahh the good old days. When you could write a schedule that nobody believed except for you and then carry on about slips and overtime. Now I have to share the same levels of fear as everybody else...

Cam pointed me at 43 things. A lot of those new web community ideas make me want to puke, but this one is strangely endearing. What is it you want to do? Just add one goal, and I bet you'll add twenty. I did.