Fixpoint

Although I had uploaded my initial release of the Gales Bitcoin Wallet components and linked them from various places, an actual article to anchor it was among the things dropped in what amounted to a mad scramble to meet MP's deadline (our initial self-imposed one being already months past). There was a not-so-hot debut, following at least in part from my not having put in much of a dress rehearsal; and at the same time the rug - or more like the whole floor, subfloor, foundation and earth beneath it - was getting pulled out from under the context in which the project had previously existed, in a progression retrospectively resembling that Hemingway line about going bankrupt: gradually and then suddenly. Then for good measure came a final spiritual blow in a stark reminder of the ever-present yet ever unexpected fragility of life and fleetingness of time.⁽ⁱ⁾

But notwithstanding these losses of original context, authority and meaning, the need that the product was designed to serve - whether perceived or not by what's left of the world - continues to exist, and that is to permit a merchant class to protect its money-holding and money-moving ability from secular power. And JWRD still exists, and works at finding, safeguarding and making the most of what scraps of sanity are still to be found in computing and even the broader environment.

So this will serve as that release article at last. Here for your convenience is a chronological index of past JWRD articles touching on GBW and its development history:

• Initial status, plan and code dump.
• Next steps in planning.
• User interface specification for two programs, an online "node" and offline "signer", with more detailed work plan.
• Code review of a historical "sendrawtransaction" patch.
• SQL database schema for a node extension to index blockchain data by address.
• Detailed presentation of Python code for the node extension: 1, 2, 3, 4, 5, 6.
• Errata discovered in my previous drafts.
• A patch for proper raw transaction support in the underlying bitcoind.
• A full walk through a rather manual sort of GBW node installation process from zero.
• An atrophied sort of introduction to Gales Scheme, the interpreter used by the offline signing program.
• Investigation and fix for a bug discovered in earlier Bitcoin Foundation work on the underlying bitcoind.
• A patch implementing predictable and adjustable transaction fee relay threshold for the Bitcoin node, among others.
• A wide-ranging examination of actual versus subverted types of Bitcoin addresses and related topics.

As MP put it,

jfw's been a busy boy, has he!

Then again, as a busy woman tells me an old saying goes,

If you want to have something done, you need to give it to a busy person.

Now as to the "re-release" part: as with my earlier work, I apparently hadn't got the habit of enclosing everything in a top-level project-name subdirectory within the version-controlled tree (perhaps because I'm just a bit too young to have suffered much through the cvs or svn days and have the git habits more ingrained - but then I wouldn't know). Yet since I've already converted most of my stuff I can't really abide keeping this one exception around; and since the legacy vpatch format never got file renaming support, it still makes most sense to just redo the genesis patch. In the same move, to firm up another convention that I now make explicit, I also put the README file directly into that project directory, as I couldn't justify to myself why it had to be buried under "package" as if it were some kind of metadata, instead emphasizing it as the foremost thing the new user is guided to and from which he is to discover his next steps.

Separated from this mechanical transformation, which applies to gbw-node and gbw-signer, are a number of packaging refinements which apply also to gscm. The most significant is the change I worked out in the log to put program links in the traditional /usr/bin rather than the djb-istic /command, thereby removing a previously implicit, non-obvious and potentially distro-dependent installation step. Thus the documented install process is now fully spelled out with nothing left to the imagination, as per how what not to do #2.

The process of obtaining the code and assembling the tree so as to have what to install in the first place, however, is not, now as then; I still have a thing upcoming for that, which as it happens is what precipitated these changes in the first place.

The present items:

Thus, the "usrbin" triad is the latest version of each and is intended to go together.

1. I guess this is the first I've said anything public about Mircea Popescu's death. I never met the man, though I would have liked to. Though I'd spent countless hours reading his output, we only spoke a few times, which I recall as being always a productive and personally tailored sort of challenge. Though I respected him, my feelings toward him were mixed; mostly, I expect, owing to my own confusions. I may have shed a tear or two, some weeks after learning of the event, as the magnitude of it gradually sunk in. At the time it was more numbness and shock. The weather was nice and I went outside and spent some hours just walking around; slowly, aimlessly, down paths in the suburban Maine woods and down unfamiliar streets, just needing time away from the routine sights, watching, feeling, and listening to the quiet world. [^]