DivNull Seeds

With the new year comes reflection and resolution. This post is a bit of both. For most of my adult life, I’ve had a tendency to start new projects, but never finish them. This begins to weigh on me, as more and more unfinished stuff accumulates. Last year, however, I realized something about the way I work, and it is going to change what I release on this blog a little bit.

Often, what excites me about a project isn’t the final product, but what I’m learning when I do it. When the learning stops, so does the whole point in doing the project in the first place. Knowing this, I’m going to start suppressing the urge to “finish” some projects before releasing them. Rather than have them sit, untouched, on my to do list forever, I’m just going to throw them out there, as is, and check them off the to-do list. I’ll be referring to projects that follow this process as “seeds”: not really done, but someone could pick them up if they want to. I’m assuming that most of these seeds will just sit there, which is fine. But, on the off chance that they provide some use to someone, I’m going to post them anyway.

I’ll try to say a bit about what I was trying to do, and what I learned with each seed I publish.

So far, here are the published DivNull seeds.

Kickstarter analysis

DivNull Productions backs a lot of Kickstarter projects. Curious about return on investment and so on, I ran the numbers on all projects DivNull backed with a “funding complete” date prior to January 1, 2012. Thus, all projects in the data set have had at least two months to deliver goals (most have had much longer). I wanted to see how much DivNull spent, what types of projects were funded, how much money went unclaimed (i.e. projects which failed to reach their funding goal) and how many projects actually delivered. Over this period, DivNull offered funding to 113 projects, 42 of which failed to reach their funding goal. This is all summarized in the following chart.

Kickstarter results through 2011

The first take away from this graph is that the number of projects not yet delivering is higher than I would have guessed. At present, 34 of the projects DivNull funded have not delivered. The monetary value obscures this a little, as around $1500 of the total can be attributed to just four of these projects. This does, however, represent just under 48% of the funded projects, which looks grim; however, the vast majority of these projects seem to be still working towards delivery.

Another surprise is that almost half of the cash offered up didn’t get used. After funding this many projects, you get a bit of a feel for which will succeed and which won’t. Initially, DivNull was fairly promiscuous about funding, taking a “give them a chance” approach, particularly when it came to role-playing games. That trailed off slightly as time went on.

As someone who gets irritated by lots of useless project updates, I tracked how many updates were made by each project. On a whim, I created a graph of the number of updates vs. the percentage of the target funds raised by the project:

Number of updates vs. Percentage of target funds

I’m not sure this reveals much of anything, except perhaps that there is no clear correlation between number of updates and funding success. And, perhaps, that overly successful projects clearly don’t update more often than other projects. I made an attempt for a while to categorize the type of comments made and count them, but ran out of interest so my dataset is incomplete. The categories may be of interest, though, especially in that they can be divided into two groups based on category: comments I care about and comments I don’t care about. All comments fell into one of these categories:

  1. “Spread the word”: requests for backers to shill for the project.
  2. “Project is progressing”: essentially a progress report without offering any new content.
  3. “Nothing is happening”: a non-progress report, usually to reassure backers that the project is still alive, even though it’s not being worked on.
  4. “Success/thanks”: Nearly every project has at least one “we did it” post.
  5. “Explaining delay or problem”: Details about why you’re not getting your swag on time.
  6. “Pitch changes/clarifications”: Adding new tiers, bonus goals provided in the event of overfunding, etc.
  7. “Pitch of related Kickstarter”: If you liked this project, you might like…
  8. “Art/content preview”: Preview text or artwork, all by itself. Sometimes art is included with other sorts of updates.
  9. “Reference to external content”: Links to a forum or blog supporting the project. Sometimes links to freebie stuff supporting the product, such as character sheets.
  10. “Incremental version”: Download to beta drafts, etc.
  11. “Request for information”: Project needs to give them information.
  12. “Release details”: Links to PDFs, details of product shipping, etc.

This list is sorted in order of least important to me to most important to me. Generally, I really only must see the last two.

Theme updated to HTML5

I finally changed this blog’s theme to some slightly more modern underpinnings, using HTML5 and CSS3. The theme rests on the foundation of the Toolbox theme from Automattic, mashed up with some bits of this site’s old theme. I also had to add in some hacky crap because Firefox hasn’t yet learned to fly right with some HTML5 tags.

While this theme looks sort of like the old one, it uses far fewer image files to get the same look. The rounded corners, drop shadows and transparent backgrounds are all done with new CSS3 capabilities (or the hacky add-ons that Firefox uses in place of them). Since not all browsers support the new hotness, you might not see either of these. Hopefully you see something that is at least legible in is place. The theme looks good appears as I intended it to on Mac versions of Safari, Chrome and Firefox and Linux versions of Firefox. I haven’t tested it on anything else.

If you spot a problem, let me know.


Now seems like a good time to provide an update on where various DivNull “products” stand, for the curious. (Feel free to add requests or questions in the comments.) Here’s what is (or, in some cases, isn’t) happening:


The closest thing DivNull has to a real software product, Omnihedron has been languishing. This is due to two main factors that happened more or less simultaneously. The first was that Wordman changed jobs into a company that avoided Windows like the plague, so access to and interest in the platform (never high) waned to zero.

At about the same time came the realization that wxWidgets, the cross-platform framework on which Omnihedron is built, wasn’t going to cut it. While it is a great framework, it has some minor issues that get amplified when applied to Omnihedron. By far the largest of these is a design philosophy which (at least at the time) implemented new features without much concern given to backward compatibility. This essentially meant that to use the framework, if you weren’t regularly keeping up with the latest changes, your code went stale. Given that development on Omnihedron tended to come in bursts separated by long gaps, this was particularly lethal. For example, on at least two occasions, when work resumed after months of stagnation, the entire system for laying out windows had been apparently deprecated, replaced with some new thing. It could be that Omnihedron just hit this sort of thing at the exact wrong time and wxWidgets is better at this now, but I don’t know.

This same issue made ports to other platforms, particularly the Mac, slow to catch up to the main branch. Worse, the Mac port was built on Carbon, an API clearly announced to be a dead end. It appears some progress has since been made on a Cocoa version, but even this is not yet ready.

All of this combined with a realization that, while DivNull has decades of professional C++ programming experience, C++ is actually a dumb programming language.

Even so, the intent is to revive Omnihedron “at some point”. A new version is certain to be Mac first, and will mostly likely be Mac only (well, maybe an iPhone version). The large stumbling block here is that DivNull doesn’t have much Cocoa experience, so will be learning.

Web Software

DivNull is working on some web software that may or may not see the light of day. Domains for these projects exist, but are being kept secret for the moment. These projects include:

  • A site dedicated to data representation in roleplaying games, built around (but not requiring) a universal scheme for storing both stats and rules. There is a “philosophy” which goes with this standard which, after a lot of thinking about it, DivNull thinks is an extremely useful way to think about such data (and also takes the standard beyond the realm of just roleplaying). This site requires a lot of documentation in order to be convincing, which doesn’t yet exist. It also opens up the possibility for about a dozen other software projects.
  • Web software (built on CakePHP) for creating “meta-indexes”. These are documents that collectively index a whole range of books, essentially mixing an index for each of the books together into one large one. The main domain idea is to gather information for a “family” of role-playing books, which often have no index at all, though other applications are possible. (This is more exciting than it sounds.) There will also be a site that hosts an instance of this software.
  • A site for posting detailed travel logs without revealing your identity to the reader. This is more of a pet project, intending to just share travel advice with friends (who would probably know who you are) in a public place, without worrying that information about, say, your family and such would link specifically to you. If you’d be interested in writing for it, let me know.


Being slightly obsessed with taking photographs of random surfaces for some time, DivNull will probably release a large collection of desktop/texture based art in some fashion, probably using Gallery 2. These will likely be released under a Creative Commons license.

Roll Play Dice Library

RollPlay exists for two reasons: 1) to define a standard way of representing dice and a parser to read it and 2) to provide much better pseudo-randomization for dice than was available at the time. The latter of these goals is now basically moot, since the standard randomization routines available now are much better than they used to be.

If RollPlay is altered (which it will need to be for Omnihedron to be moved to Cocoa), it will likely be by adding support for other languages, rather than changing much else about it.

And no, it still doesn’t need to support parentheses.


Most roleplaying work continues to focus on Exalted. A collaborative netbook of Exalted artifacts called When Autochthon Dreams will be released by the end of the year.

Even while this work continues, DivNull’s disaffection with Exalted increases. Unless this changes, the collected charm tree project will finish out Second Edition, but will not continue beyond that.

End of life

Titles marked as “final version” on the software page will not be changed or upgraded, but will still be available for download.


After eight years, the main DivNull site is finally getting a redesign. In addition to a new look, the site is now backed by the leading blogging software. Most importantly, it now supports RSS feeds.

In doing this redesign, it was also decided to change names from DivNull Software to DivNull Productions, and to merge the various non-software production we have done all along under the main label. This was done mostly to provide “one-stop shopping” for people interested in what we are making these days, but also because the line between what is “software” and “not software” is blurring so much that the distinction is loosing its meaning.

As much of our “not software” work consisted of products for tabletop role-playing games, expect to see more news of such on this page from now on. Announcements will be categorized (see the categories to the right).

We are also actively working on some web-based projects that may or may not see the light of day. If they do, they will likely debut under their own domain names. Even so, this site will still be used to make announcements about them.

Thanks to those of you who have supported DivNull over the years. We hope to make the coming years even better.

The DivNull Pledge

DivNull makes the following assertions about its software:

  • No DivNull product will ever use a folder tab control.
  • DivNull code will be free from divide by zero errors.
  • DivNull titles will never use the words “Super”, “Hyper”, “Pro”, “Plus”, “Power”, “X”, or the prefix “i” in their names.