World Machine Development Diary

Hi folks,

I just realized that its been a while since I last posted here, so I figured I had best make an  update!

There’s been lots of work being done but not much visible progress lately. Partially this is because I’ve been doing a fair bit of work not directly related to the main  World Machine codebase  lately. However, more important is the fact that I’ve been doing lots of work in the last few months or so in improving the error handling of World Machine.

WM never really has handled error conditions particularly well; your answer to asking to do something beyond its capabilities was almost always a crash. The crash reporting handler in the last beta was the first step in refining things. For this next beta, enormous amounts of work have been put in to make World Machine maintain stability, recover gracefully, and inform you properly when things don’t go according to plan. This is particularly true of memory handling; My goal for the next beta is that you will be unable to make WM crash no matter what you ask of it! (Whether this is achieved will remain to be seen..)

Hi folks,

It’s been a while so figured I should write a little update! I hope all of you are having a great holiday. Things have been extremely hectic here both personally and professionally over the winter to date so I’ve been pretty bad at keeping up with my correspondence — if you’ve emailed me and not received a response yet, my apologies.

There are some interesting changes coming down the pipe in the next 2.3 Beta due to some contracted changes requested by a customer that will be shared to everyone. More about that in the next post. But for now I just wanted to get a shout out that I am alive and wish you all the best this season!

Hi guys,

Just wanted to post this image which I thought demonstrated a use of mask outputs that I haven’t seen much of yet.

I’ve created a basic terrain, eroded and then placed a road into the terrain via a spline path.

What’s interesting, is that I’ve then looked at the terrain before and after the road insertion to determine the areas that have been cut, and the areas filled, and then textured them appropriately:

Capturing cuts and fills

What struck me how simple this information was to extract from the WM network, but how bloody useful it is to anyone doing paths on terrains!

Been working on fixing numerous display bugs lately — one of them turned out to actually be an ATI display driver issue that was fixed in an update, the other another strange driver quirk with a viable workaround.

Graphics compatibility vs performance has always been a challenge, as a content tool WM is sometimes run on substandard hardware and needs to remain compatible with that… while at the same time, when displaying high resolution output requires real-time performance as well.

I’m thinking WM 2.3 Beta-2 will be released sometime in the next week or so to bump out the changes to everyone!

Beta-2?

Beta-2 is accumulating a nice set of bugfixes, most of which were NOT introduced in Beta-1 which is nice. I have had an isolated report of a crash issue in B1 which I have not had replicated or further information about.

Layout View

I’m right now going through and fixing some minor irritating issues with the Layout View, including inconsistencies in the UI and a couple other issues. I also would like to add a few more tools into the layout toolset — certainly allowing polygons to have custom heights set at any knot is a highly requested feature.

Anyone have any other wishlists I should know about?

Beta 1 Available! Rejoice!

As I mentioned on the forums, Beta-1 of WM 2.3 is now available if you haven’t heard! Visit the main website and follow the news links to get at it.

I’ve already started fixing a couple further bugs, and introducing a few new things for Beta-2, such as multithreaded support for Thermal Erosion.

I’ve been continually mentioning how 2.3 is mostly an internals change that should pave the way for easier and more profound changes down the line. I’m hoping to be able to give folks a taste of those changes sooner rather than later.

What I really want to talk about is the concept of Code Debt.

The curse of Code Debt

One experience I’ve had on the World Machine project is just how toxic the kind of “code debt” that 2.3 addresses is. For those who aren’t familiar with the concept, Code Debt is essentially much like financial debt.

When you need something done in a whole hurry, and you don’t have the resources available, you can choose to “take out a loan” by coding things in the fastest but not best engineered way. In short, a dirty hack.

However, you have to pay interest on that loan. And the interest rate is high : Every time you need to touch that system in the future, you have to interact with your crude code. Over time, the cost of having to deal with your dirty hack far outweighs the time savings you got in the first place.

All that Work, but still the Same?

To deal with the code debt, you have to continually refactor your code to improve its quality. WM 2.3 is primarily a release aimed at doing just this. Lots of internal systems are taken apart, re-plumbed, and put back together again. The depressing thing though, is when you’re done..

… you have exactly what you started with, from the outside. In the ideal case, nothing will have changed from the User’s perspective. If you get alot of your development drive out of creating cool new features, like I do, this can be a bit depressing.

The good news of the effort though, is that done well you can move forward from that spot at a far faster clip than would have been possible if the debt was not dealt with.

Anyways, just a kind of windy aside about what the primary goals of WM 2.3 were. I hope it provides some illumination for folks out there!

Hey folks,

I had originally planned to release the first beta of WM 2.3 today, but I realized only yesterday(!) that that is probably not a good idea since I’m going out of town this weekend to help some friends out — in case anything is horribly broken I wouldn’t be able to fix it in realtime.

So I’ve postponed the 2.3 beta again, until monday, so that I can be sure that I will be immediately able to fix anything that arises.

See you then!

A good thing I pushed 2.3 back a little bit further, as I’m currently fixing a serious build crash that was uncovered once I started the final testing. And let me tell you, there’s nothing like tracking down subtle multi-thread crash bugs to really brighten ones day!

I also want to fix the “stranded route points” issue for 2.3beta-1.

I intended to actually open up 2.3 Beta1 tonight to all registered customers, but I’ve decided to push that back a bit longer.

The main impediment is actually not the product itself but all of the supporting web code: I wanted to create a customer service center where registered users (and ONLY registered users..) can always go to download the latest stable and beta editions for their purchases, etc. This will ease the deployment issues I face for doing many smaller releases rather than mammoth ones. Probably about a week worth of work to get that up and running next week.

Long story short: If you’re a WM 2.2 user who urgently wants to get at the bugfixes from 2.3Beta-1, just drop me a line at stephen@world-machine.com and I’ll make sure you get 2.3B1 immediately! Otherwise it should roll out next week.

Cheers!