@@ -1,61 +1,68 @@
11+++
22title = "Toward a New %clay"
3−date = 2016-07-14
4−author = "Curtis Yarvin"
3+date = "2016-07-13"
54description = "Urbit's revision-control system, %clay, is itself due for a (medium-sized) revision!"
6−[taxonomies]
7−posts = ["Essays"]
5+aliases = [
6+ "/posts/essays/toward-a-new-clay",
7+ "/posts/toward-a-new-clay",
8+ "/blog/clay"
9+]
10+
11+[extra]
12+author = "Curtis Yarvin"
13+ship = "~sorreg-namtyv"
814+++
15+
916TLDR: (a) Urbit's revision-control system, `%clay`, is itself due
10−for a (medium-sized) revision! (b) this time, you yourself can
11−really help! (c) please don't give away the buried spoiler!
17+for a (medium-sized) revision! (b) this time, you yourself can
18+really help! (c) please don't give away the buried spoiler!
1219
1320This explanation does not (or at least should not) require you to
1421know anything about Urbit.
1522
1623### But %clay is already pretty cool
1724
18−Of course, nothing in Urbit is allowed to be _lame_. `%clay` is
19−actually quite cool. It does need some work. But first, let's
25+Of course, nothing in Urbit is allowed to be _lame_. `%clay` is
26+actually quite cool. It does need some work. But first, let's
2027talk about what it is and why it's cool.
2128
2229`%clay` is the Arvo vane (Urbit kernel module) for revision
23−control. It's an authenticated, human-memorable global immutable
24−namespace which is also a typed revision-control system. It's
25−also very good at subscription and synchronization. Urbit
30+control. It's an authenticated, human-memorable global immutable
31+namespace which is also a typed revision-control system. It's
32+also very good at subscription and synchronization. Urbit
2633currently uses `%clay` to distribute live hotpatches; we
2734transparently upgrade the whole network with one checkin.
2835
2936### Compared to other revision-control systems
3037
31−`%clay` is clearly a vanilla DVCS in the modern `git` style. Of
38+`%clay` is clearly a vanilla DVCS in the modern `git` style. Of
3239deployed DVCS these days, `%clay` is probably closest in spirit
3340to `bzr` or maybe `fossil`.
3441
3542Obviously, as a part of a larger whole, you would expect any
36−Urbit RCS to be brutally spartan. `%clay` is. It's even more
43+Urbit RCS to be brutally spartan. `%clay` is. It's even more
3744spartan than `bzr`.
3845
39−(In fact, it's really too spartan to fully use at the moment. We
46+(In fact, it's really too spartan to fully use at the moment. We
4047ourselves still use `%clay` mainly as a distribution system, and
41−do all internal development on Github. This is because the
48+do all internal development on GitHub. This is because the
4249present `%clay` lacks a few little conveniences, like commit
43−messages. See upgrades, below.)
50+messages. See upgrades, below.)
4451
45−There are lots of normal DVCS things in `%clay`. But there are
46−also cool and unusual ones. Let's take a brief tour of the
52+There are lots of normal DVCS things in `%clay`. But there are
53+also cool and unusual ones. Let's take a brief tour of the
4754high-level system, then describe what we think should change
4855about it.
4956
5057### Anatomy of a DVCS
5158
52−A DVCS is a historical state graph. A point in this graph is a
59+A DVCS is a historical state graph. A point in this graph is a
5360snapshot of some tree-structured system, like a filesystem.
5461
5562These snapshots, including recursive internal nodes, are stored in
56−a content-addressable store. Pointers in this store are used to
63+a content-addressable store. Pointers in this store are used to
5764construct a revision graph, whose edges are edit and/or merge
58−actions. Some sort of namespace attaches meaningful names to
65+actions. Some sort of namespace attaches meaningful names to
5966current revisions in this graph.
6067
6168There are six big questions in the design of a DVCS:
@@ -71,16 +78,16 @@ There are six big questions in the design of a DVCS:
7178
7279A good way to summarize `%clay` (at present) is that its answers
7380to the first four questions are interesting, and its answers to
74−the last two are boring. We'll talk about all six, though.
81+the last two are boring. We'll talk about all six, though.
7582
7683### Global namespace: [ship desk case]
7784
78−The permanent global name of a `%clay` snapshot is a `++beak`. A beak
85+The permanent global name of a `%clay` snapshot is a `++beak`. A beak
7986is a triple: `[ship desk case]`.
8087
81−The `++ship` is an Urbit ship. A ship is a network address, a
82−human-memorable name and a cryptographic identity. The address
83−width is variable; shorter addresses are more desirable. Most
88+The `++ship` is an Urbit ship. A ship is a network address, a
89+human-memorable name and a cryptographic identity. The address
90+width is variable; shorter addresses are more desirable. Most
8491users have 32-bit addresses, which we map to four-syllable
8592synthetic names like `~sorreg-namtyv`.
8693
@@ -93,44 +100,44 @@ In the Urbit path syntax, the beak is the first three segments of
93100the path: `/ship/desk/case/my/cat/picture`.
94101
95102In most cases, a path parser has a default ship, desk and case,
96−each of which can be replaced by `=`. So `/=home=/my/cat/picture`
103+each of which can be replaced by `=`. So `/=home=/my/cat/picture`
97104is the data at `/my/cat/picture` on your own ship, in the `%home`
98−desk, right now. Immutability doesn't have to be inconvenient.
105+desk, right now. Immutability doesn't have to be inconvenient.
99106
100−`%clay` has no concept of a mutable name or a magic `HEAD`. Use
101−the current date. If you're referring to remote data and you're
107+`%clay` has no concept of a mutable name or a magic `HEAD`. Use
108+the current date. If you're referring to remote data and you're
102109worried about clock skew, subtract a second or three, or be
103−prepared to block. (Or keep your clock unskewed, which shouldn't
104−be _that_ hard.) It's not like mutable-name schemes work any
110+prepared to block. (Or keep your clock unskewed, which shouldn't
111+be _that_ hard.) It's not like mutable-name schemes work any
105112better than this anyway.
106113
107−(Immutability is central to `%clay`. `git` merely adopted the
108−reflog. `%clay` was born in it, molded by it...)
114+(Immutability is central to `%clay`. `git` merely adopted the
115+reflog. `%clay` was born in it, molded by it...)
109116
110117`%clay` is objective; each Urbit node hosts only one ship, but
111−treats domestic and foreign data (almost) the same way. If you
118+treats domestic and foreign data (almost) the same way. If you
112119ask your `%clay` repository for someone else's data, it will
113120block the request until it gets the data, which it then caches.
114121At present, we always fulfill requests from the origin ship,
115122though in future we could certainly use it as a tracker.
116123
117124If `%clay` discards cached data, it keeps a hash, so that your
118−urbit can't violate the local illusion of immutability. We
125+urbit can't violate the local illusion of immutability. We
119126can't, of course, prevent anyone else's ship from sending
120−conflicting claims in different directions. But we can prevent
127+conflicting claims in different directions. But we can prevent
121128programs on our own ship from hearing about them, meaning that we
122129retain local referential transparency.
123130
124131### Local structure: dome, arch, leaf and mark
125132
126−A `%clay` snapshot is called a `++dome`. A dome is the `%clay`
133+A `%clay` snapshot is called a `++dome`. A dome is the `%clay`
127134equivalent of a tree-structured filesystem.
128135
129136Each node in this tree is one uniform structure, an `++arch`,
130−which is both file and directory. A `++leaf` or data value in
137+which is both file and directory. A `++leaf` or data value in
131138this arch is a typed document.
132139
133−A logical document type is called a `++mark`. A mark is a
140+A logical document type is called a `++mark`. A mark is a
134141symbolic pointer to a Hoon source file, in the same dome, for an
135142engine that can validate, compare, and/or convert
136143documents of that mark.
@@ -139,8 +146,8 @@ Let's go over arches and marks a little more closely...
139146
140147#### An arch is a uniform filesystem node
141148
142−Most filesystems and RCS have modal nodes. For instance, a git
143−data node can be a blob or a tree. `%clay` (and the Urbit
149+Most filesystems and RCS have modal nodes. For instance, a git
150+data node can be a blob or a tree. `%clay` (and the Urbit
144151namespace in general) has one kind of node, an `arch`:
145152
146153```
@@ -149,24 +156,24 @@ namespace in general) has one kind of node, an `arch`:
149156```
150157
151158An arch has an optional typed-data leaf, and optional named
152−children. In Unix terms, it's a file, a directory, or both. It
159+children. In Unix terms, it's a file, a directory, or both. It
153160can't be neither: an empty arch, with no data and no children, is
154−auto-pruned. As in `git` (but not in Unix), the hierarchy is a
155−pure function of its leaves. Paths exist only because of the
161+auto-pruned. As in `git` (but not in Unix), the hierarchy is a
162+pure function of its leaves. Paths exist only because of the
156163data they point to.
157164
158165The `++knot` names in an arch path are also a little special.
159166They're typed, in a sense; any knot can be parsed invertibly into
160−a typed atom. The atom is an arbitrary unsigned integer. Its
167+a typed atom. The atom is an arbitrary unsigned integer. Its
161168type (an `++aura` if you know Hoon) defines the semantics and
162169presentation of the atom: for instance, a symbol, a decimal,
163170a date, etc.
164171
165172Ultimately a filesystem is a database, and a path is a primary
166−key. It's very useful to be able to store simple data in this
173+key. It's very useful to be able to store simple data in this
167174key, in a standard and unambiguous way.
168175
169−Why uniform nodes? One, they map nicely to information trees
176+Why uniform nodes? One, they map nicely to information trees
170177where a subtree may need some custom collective presentation,
171178not an automatically generated Apache style directory listing.
172179Two, non-uniform nodes introduce a whole pointless category of
@@ -174,34 +181,34 @@ node type conflicts -- for which life is far too short.
174181
175182#### A mark is an executable document type
176183
177−A mark is "like an executable MIME type." It's best defined as a
184+A mark is "like an executable MIME type." It's best defined as a
178185symbolic mapping from a filesystem to a schema engine.
179186
180187If you have a MIME type `text/html`, you need to watch the
181−Internets to track what the W3C thinks HTML is. If you have an
188+Internets to track what the W3C thinks HTML is. If you have an
182189Urbit mark `%html`, you can functionally construct (using our
183190insanely powerful functional build vane `%ford`) a core which
184191can validate or normalize HTML, apply patch and diff logic, and
185−translate to and from other marks. `%ford` can even reason about
192+translate to and from other marks. `%ford` can even reason about
186193translation paths and construct multistep conversions.
187194
188195(Urbit does not actually have this super-smart `%html` mark right
189−now. But it probably should, huh. And this page was composed in
196+now. But it probably should, huh. And this page was composed in
190197Markdown and translated to HTML by `%ford`.)
191198
192−Where does the source code of a mark live? Within the same dome
193−as the data it describes. `%html` files are described by the
199+Where does the source code of a mark live? Within the same dome
200+as the data it describes. `%html` files are described by the
194201Hoon source at `/===/mar/html`, which perhaps gets mapped into
195202your Unix filesystem at `$ship/home/mar/html.hoon`.
196203
197−This Hoon file is not self-contained. It has dependencies. The
204+This Hoon file is not self-contained. It has dependencies. The
198205`%html` mark may depend on various structures and libraries
199−elsewhere in Urbit. But the source code for all them is within
206+elsewhere in Urbit. But the source code for all them is within
200207this dome.
201208
202209One constraint this puts on domes is that _all domes are
203−descendants of the base system dome_. Domes are versions of one
204−reality, not fragments of that reality. You can't create a new
210+descendants of the base system dome_. Domes are versions of one
211+reality, not fragments of that reality. You can't create a new
205212empty dome -- you'd have no mark source, and be unable to install
206213any validated files.
207214
@@ -212,13 +219,13 @@ weird collisions between non-parallel things don't happen.
212219
213220### Unix mapping and mounting
214221
215−Every DVCS is its own filesystem in a sense. But every DVCS
222+Every DVCS is its own filesystem in a sense. But every DVCS
216223works by mapping and mounting itself, in some sense, to Unix.
217224
218225#### Mapping arches to Unix
219226
220−How do we map arches into Unix? The arch structure maps 1:1 to
221−the Unix directory tree. When the arch has a leaf, we derive a
227+How do we map arches into Unix? The arch structure maps 1:1 to
228+the Unix directory tree. When the arch has a leaf, we derive a
222229dot extension from the file's mark.
223230
224231So, supposing we have a leaf `foo` which is an HTML document
@@ -232,19 +239,19 @@ foo/moo.html
232239```
233240
234241This maps invertibly back into `%clay`, and it looks perfectly
235−unremarkable from the Unix side. File completion sometimes gets
242+unremarkable from the Unix side. File completion sometimes gets
236243a little annoying.
237244
238245It would be if nice every possible Unix tree, even every
239246plausible Unix tree, mapped cleanly and completely to a `%clay`
240−dome. There are always horrid exceptions, even in common use --
241−like extensionless Makefiles. The trivial solution is that if
247+dome. There are always horrid exceptions, even in common use --
248+like extensionless Makefiles. The trivial solution is that if
242249`%clay` finds a file mysterious, it won't track it.
243250
244251One often unquestioned assumption in new system software is that
245−compatibility with old content is important. It's nice, but not
246−that important. Most content in a new system is new content.
247−The important compatibility direction is export, not import. So
252+compatibility with old content is important. It's nice, but not
253+that important. Most content in a new system is new content.
254+The important compatibility direction is export, not import. So
248255the new system should subset, not superset, the ancient formats.
249256
250257### Mounting desks to Unix
@@ -255,36 +262,36 @@ the Unix directory of the same name within the pier.
255262When files in a mount directory change, `%clay` notices it
256263through `libuv` filesystem watching (ultimately `inotify()` or
257264`FSEvents`), computes diffs, and automatically commits them to
258−the desk. And if `%clay` creates a new version for its own
265+the desk. And if `%clay` creates a new version for its own
259266urbity reasons, it will revise the Unix files.
260267
261268Keeping all mounted files in the same Unix subdirectory makes it
262−easy to move the whole pier in a tarfile. Edits applied while
269+easy to move the whole pier in a tarfile. Edits applied while
263270the `urbit` process isn't running will be detected and applied
264271the next time it starts.
265272
266273In conventional DVCS terms, `%clay` has a 1:1 static mapping from
267−branches to working directories. Why?
274+branches to working directories. Why?
268275
269−Why not? Do we need that "what branches are checked out in what
270−directories" piece of state? Why force the user to worry about
271−which branches are mounted to which directories? Why not just
272−mount them under their own names? Is there some crippling global
276+Why not? Do we need that "what branches are checked out in what
277+directories" piece of state? Why force the user to worry about
278+which branches are mounted to which directories? Why not just
279+mount them under their own names? Is there some crippling global
273280shortage of disk space for source files? Our planet is running
274−out of inodes? Wut?
281+out of inodes? Wut?
275282
276283### Queries, subscriptions and synchronization
277284
278285There are two ways to get data out of `%clay`: dereference the Urbit
279−namespace, or send a request/subscription. Sync is in userspace.
286+namespace, or send a request/subscription. Sync is in userspace.
280287
281288#### Dereferencing the namespace
282289
283−`%clay` is just part of the full Urbit namespace. The full namespace
290+`%clay` is just part of the full Urbit namespace. The full namespace
284291has a two-character prefix which expresses what vane to query
285−(always `%c` for `%clay`). The suffix is a mode flag; `%x` for
292+(always `%c` for `%clay`). The suffix is a mode flag; `%x` for
286293data, `%y` for a list of children, `%z` for the whole arch
287−subtree, `%w` for the version number. (Each vane has its own
294+subtree, `%w` for the version number. (Each vane has its own
288295modes.)
289296
290297For example, the full path to my cat picture is
@@ -298,6 +305,7 @@ For example, the full path to my cat picture is
298305If the data model is `++image` (hypothetical -- we don't really
299306do image processing in Urbit right now), the normal way to load
300307this data in a Hoon program is
308+
301309```
302310.^(image /cx/~sorreg-namtyv/home/[now]/my/cat/picture)
303311```
@@ -314,15 +322,15 @@ learn it in future, `[~ ~]` if we know we'll never know it, or
314322`[~ ~ value]` if we know it now.
315323
316324In high-level shell programming (`:dojo`), we can make apparently
317−"blocking" requests to `%clay`. What actually happens: when `%clay`
325+"blocking" requests to `%clay`. What actually happens: when `%clay`
318326produces `~`, we stop trying to compute this expression and wait
319−on one or more blocking requests. When we are unblocked, we
327+on one or more blocking requests. When we are unblocked, we
320328recompute again from scratch; the intermediate computation is not
321329saved at all.
322330
323−Urbit doesn't like to block. It likes to save coherent data
324−structures, not incomplete computation contexts. When Urbit
325−pretends to block for you, this is a courtesy service. You
331+Urbit doesn't like to block. It likes to save coherent data
332+structures, not incomplete computation contexts. When Urbit
333+pretends to block for you, this is a courtesy service. You
326334can return the courtesy of not expecting miracles from it.
327335When in doubt, send an actual request.
328336
@@ -333,7 +341,7 @@ namespace, but can (a) read across case (revision) ranges,
333341and (b) filter by subtree.
334342
335343If the revision range in a request stretches into the future, the
336−request becomes a subscription. Urbit messaging always regards a
344+request becomes a subscription. Urbit messaging always regards a
337345single response as a special case of a subscription.
338346
339347Generally, the response stream created by a request starts with
@@ -343,10 +351,10 @@ closed when the request range is complete.
343351
344352#### Userspace synchronization
345353
346−Synchronization per se is done in userspace. In `%clay`, sync just
354+Synchronization per se is done in userspace. In `%clay`, sync just
347355means an application (typically the `%kiln` module in the system
348356app `:hood`) subscribes to a remote desk, and saves its change
349−stream locally. So sync is really an emergent feature of `%clay`.
357+stream locally. So sync is really an emergent feature of `%clay`.
350358
351359### Revision graph
352360
@@ -354,16 +362,16 @@ stream locally. So sync is really an emergent feature of `%clay`.
354362Some changes are proposed below, however.
355363
356364As in any DVCS, a dome is the child of one parent dome (an edit)
357−or more parents (a merge). Often the parent of any revision is
365+or more parents (a merge). Often the parent of any revision is
358366the previous version in the sequence, but it can be anything.
359367
360368### Storage model
361369
362−`%clay`'s storage model is completely uninteresting. It's also a
370+`%clay`'s storage model is completely uninteresting. It's also a
363371little funky and I want to change it, so I'll be super vague.
364372
365373A `++lobe`, a 128-bit atom, is the Merkle hash of a `blob`, which
366−is essentially a leaf. All blobs are stored in a hash table
374+is essentially a leaf. All blobs are stored in a hash table
367375(`(map lobe blob)`), and reference each other through lobes.
368376
369377At present we don't GC this space at all, because we don't
@@ -372,12 +380,12 @@ actually delete any roots.
372380A `blob` may be `%direct`, meaning it's defined as the data
373381itself, or `%delta`, defined as a change against another blob.
374382This is motivated by the desire not to store a full snapshot
375−on every file change. But it may be misguided (see changes #0
383+on every file change. But it may be misguided (see changes #0
376384and #1 below).
377385
378386We cache the full arch tree (`++ankh`) of the latest state in
379−each desk. For current queries, this mitigates a lot of the
380−opportunities for slowness that this data model creates. But the
387+each desk. For current queries, this mitigates a lot of the
388+opportunities for slowness that this data model creates. But the
381389disparity between the speed of the current version and the
382390speed previous version is an undesirable glitch in the API.
383391
@@ -389,38 +397,38 @@ We actually have _three_ motivations for change...
389397
390398At present, we're still using `git` as a development platform.
391399`%clay` is certainly used, but just for its strength --
392−deployment. This is starving Urbit of essential experience and
393−basically keeping it lame. So, change is necessary.
400+deployment. This is starving Urbit of essential experience and
401+basically keeping it lame. So, change is necessary.
394402
395403### Change must come from the community
396404
397−"Why?" you ask? "Why must change come from the community?" Well,
398−one, duh. Change always comes from the community.
405+"Why?" you ask? "Why must change come from the community?" Well,
406+one, duh. Change always comes from the community.
399407
400−But two, a great force for change has departed. Philip Monk,
408+But two, a great force for change has departed. Philip Monk,
401409late maintainer of `%clay` (`~wictuc-folrex` in the old era,
402−`~wicdev-wisryt` in the new), is tired of types and typing. He
410+`~wicdev-wisryt` in the new), is tired of types and typing. He
403411spent September roaming with the bison in Montana, and is now
404−literally on his way to Patagonia. On the one hand, Phil still
412+literally on his way to Patagonia. On the one hand, Phil still
405413exists and is not at all unwilling to answer questions, even dumb
406−questions. On the other hand, _Patagonia_.
414+questions. On the other hand, _Patagonia_.
407415
408−So %clay needs a new owner, or owners. At this point, we think
416+So %clay needs a new owner, or owners. At this point, we think
409417the community might be indoctrinated enough to take a stab.
410418
411419Alas, `%clay` already has a pretty checkered ownership history.
412420Long ago I built a `%clay` that was really the only serious
413−design failure in Arvo. Then Jared Hance, a brilliant summer
421+design failure in Arvo. Then Jared Hance, a brilliant summer
414422intern, took my broken `%clay` and reformed it into a minimal `git`
415−in about a week. Then Philip modified Jared's `%clay` to be a
423+in about a week. Then Philip modified Jared's `%clay` to be a
416424_typed_ revision-control system, which is a pretty cool thing.
417425
418426But... in none of these transitions did `%clay` actually get
419−_rewritten_ per se. While it's a reasonably good vane, it
427+_rewritten_ per se. While it's a reasonably good vane, it
420428does contain one or two historical decisions.
421429
422430We know a lot more about Arvo than we used to, and I think we can
423−revisit many of these decisions. The point of this post is to
431+revisit many of these decisions. The point of this post is to
424432start a `%clay` design discussion that will evolve into actual
425433collective action.
426434
@@ -430,41 +438,41 @@ We're planning a rewards program for contributions to `%clay`.
430438It'll simply involve setting a time period and a pool of real
431439estate, and allocating the pool by judged contribution.
432440
433−This is an experimental program. We'll say more about it later.
434−It may not even happen. It may happen very differently. When in
441+This is an experimental program. We'll say more about it later.
442+It may not even happen. It may happen very differently. When in
435443doubt, don't talk about Fight Club.
436444
437445## Some openings for change
438446
439−Change, of course, must come from the community. No one can
440−dictate change from the top down. We _know_ this. But...
447+Change, of course, must come from the community. No one can
448+dictate change from the top down. We _know_ this. But...
441449
442450This list of suggested changes for the next `%clay` is not in any
443−sort of order: priority, structure, difficulty, etc. Let's talk
444−about the changes first; make actual plans second. The numbers
451+sort of order: priority, structure, difficulty, etc. Let's talk
452+about the changes first; make actual plans second. The numbers
445453here are just for convenience.
446454
447−*All changes in this program will happen on the next-generation
448−(`cc-release`) branch.*
455+_All changes in this program will happen on the next-generation
456+(`cc-release`) branch._
449457
450458### Change #0: state reform
451459
452460As mentioned above, the storage model in `%clay`, as the code is
453−at present, feels a little funky to me. There's nothing terribly
461+at present, feels a little funky to me. There's nothing terribly
454462awful about it, but it doesn't feel quite right and its scaling
455463properties are very questionable.
456464
457465Briefly, `%clay` today actually derives the hierarchy from its
458−leaves -- rather than _logically_ so deriving it. Only leaves
459−are in the object store. This is just weird, and you can feel
460−`%clay` trying to be something other than what it is. (I blame
466+leaves -- rather than _logically_ so deriving it. Only leaves
467+are in the object store. This is just weird, and you can feel
468+`%clay` trying to be something other than what it is. (I blame
461469Jared, though it's actually my fault.)
462470
463471To store a dome snapshot (`++yaki`), the present code keeps a
464−`(map path blob)` at the root of the dome. This is enough to
472+`(map path blob)` at the root of the dome. This is enough to
465473reconstruct the `++arch` structures the namespace exports.
466474However, all kinds of basic arch operations, like basically `ls`,
467−become `O(n)` in the number of leaves in the dome. This is
475+become `O(n)` in the number of leaves in the dome. This is
468476obviously not quite right.
469477
470478I believe the storage structure should be (removing everything
@@ -485,46 +493,46 @@ Here a `++root` is the universal node, with children addressed by
485493hash, but logically representing the abstract `++arch`.
486494
487495Users of hierarchical filesystems basically expect performance
488−qualities that match the filesystem structure. A huge directory
489−in one place shouldn't affect search speeds somewhere else. And
496+qualities that match the filesystem structure. A huge directory
497+in one place shouldn't affect search speeds somewhere else. And
490498of course, `ls` can't be O(n) in the size of the filesystem.
491499
492500### Change #1: store edges, cache snapshots
493501
494−The current `%clay` is snapshot-oriented, like `git`. The actual
495−changes are discarded. I think this is a bad choice.
502+The current `%clay` is snapshot-oriented, like `git`. The actual
503+changes are discarded. I think this is a bad choice.
496504
497505`%clay` is not a research project, and has no idea of going as
498−far as `darcs` in the "set-of-patches" direction. However, at
506+far as `darcs` in the "set-of-patches" direction. However, at
499507present we actually throw away the patch (`++nori`).
500508
501509I think the right DVCS design for `%clay` is a snapshot-oriented
502−system which also keeps the patches. The revision graph has the
503−same general shape as in the old `git` style. But rather than
510+system which also keeps the patches. The revision graph has the
511+same general shape as in the old `git` style. But rather than
504512just deriving the diffs from the snapshots, we can also derive
505−the snapshots from the diffs. When in doubt, we save space by
513+the snapshots from the diffs. When in doubt, we save space by
506514dropping the snapshot, not the diff.
507515
508516For an RCS which is also a general-purpose typed filesystem, it's
509517clear that considering edges (deltas, diffs, changes, etc) as the
510−primary state is optimal. You can use diff functions to derive
518+primary state is optimal. You can use diff functions to derive
511519deltas from snapshots, but in many data models this will always
512−risk losing information from the original action. When in
520+risk losing information from the original action. When in
513521doubt, save what the user did.
514522
515523This is even more true for operations on the tree structure, not
516−on individual leaves. Consider the heuristic energy that `git` has
524+on individual leaves. Consider the heuristic energy that `git` has
517525to go through to figure out when `mv` has been used -- even when
518−`git mv` has been used! (According to the `git` faq, this is
526+`git mv` has been used! (According to the `git` faq, this is
519527because the lowest common denominator remains Linus' style of
520528working with emailed patches.)
521529
522−Edges probably don't even need to be deduplicated. They also can
530+Edges probably don't even need to be deduplicated. They also can
523531contain metadata such as commit messages (which `%clay` has no
524532place for at present).
525533
526534Once edges are recognized as unreclaimed primary state, snapshots
527−can always be derived from the edges. So we can reclaim memory
535+can always be derived from the edges. So we can reclaim memory
528536being used for snapshots.
529537
530538It's also straightforward to have a kind of edge which is a list
@@ -532,27 +540,27 @@ of edges, letting us collapse feature branches trivially.
532540
533541Finally, I think references in edges should wherever possible
534542indirect through the namespace, rather than directly to the
535−content address. The user has strictly more information if you
543+content address. The user has strictly more information if you
536544say you merged `/~sorreg-namtyv/base/987/`, than if all you have
537545is the hash that this reference refers to.
538546
539547### Change #2: unconfuse merge and rebase
540548
541549The dispute between merge and rebase has all the bad smells of
542−a historical wart. But here I'm just not sure what's right.
550+a historical wart. But here I'm just not sure what's right.
543551
544552The user has one _intent_ whether merging or rebasing, which is
545−to combine changes on one desk into another. The _algorithm_
553+to combine changes on one desk into another. The _algorithm_
546554used to construct this change, whether operational transformation
547555(rebasing) or a 3-way diff (merging) has one goal: use all
548556available information to model the authors' editing intent as
549−realistically as possible. If we save edges, we should use
557+realistically as possible. If we save edges, we should use
550558edges.
551559
552560Whether to define the resulting change graph with the typical
553561structure of a merge, or the typical structure of a rebase,
554562strikes me as quite arbitrary, and well into the land of "pick
555−one and force everyone to stick with it." But this call demands
563+one and force everyone to stick with it." But this call demands
556564a real expert on RCS workflow, which I'm not.
557565
558566Since `%clay` is a humble part of a larger system and needs to
@@ -562,23 +570,23 @@ different preferences" is not usually the right answer.
562570### Change #3: comment jihad
563571
564572`%clay` needs to be brought up to the documentation standards of
565−`%jael`. This means (a) the formatting matches; (b) the comments
573+`%jael`. This means (a) the formatting matches; (b) the comments
566574are at the same level of quality and ubiquity.
567575
568576The basic commenting standards are: (a) every core (class) needs
569577a block comment explanation; (b) every gate (function) needs a
570578line comment; (c) every face (variable) needs a line comment.
571579
572−This is obviously not just a clerical task. To write these
573−comments, you need to actually understand the code. This is a
580+This is obviously not just a clerical task. To write these
581+comments, you need to actually understand the code. This is a
574582nontrivial but extremely rewarding task, even if the code you're
575583commenting will get replaced.
576584
577585Also, the _symbols_ in `%clay` must be brought up to `%jael`
578−quality. There is still far too much four-letterism here.
586+quality. There is still far too much four-letterism here.
579587
580588NB: `%clay` is about 3000 lines of pretty clean Hoon, which is
581−certainly not at all undocumented. We're not exactly looking at
589+certainly not at all undocumented. We're not exactly looking at
582590a Matterhorn of reverse engineering here.
583591
584592### Change #4: knot to coin internal paths
@@ -589,26 +597,26 @@ ASCII symbol as an atom) must be within the range of the
589597invertible noun printer, `++co`.
590598
591599`++co` maps a described data value, a `++coin`, to an ASCII
592−string which is a valid URL segment and Unix filename. Its
600+string which is a valid URL segment and Unix filename. Its
593601friend `++so` inverts the function, perfectly.
594602
595−Almost every knot is the simplest kind of `coin`, a `dime.` A
603+Almost every knot is the simplest kind of `coin`, a `dime.` A
596604dime is a single atom with a rendering type (`aura` if you know
597−Hoon). Auras include `@tas` for symbolic text, `@ud` for
598−unsigned decimal, `@p` for Urbit ship, etc. `++co` can also
605+Hoon). Auras include `@tas` for symbolic text, `@ud` for
606+unsigned decimal, `@p` for Urbit ship, etc. `++co` can also
599607print arbitrary cells, though less prettily.
600608
601609This feature is cool because it lets us stuff data into names in
602−a precise and predictable way. Usually the names in a namespace
610+a precise and predictable way. Usually the names in a namespace
603611are human-meaningful and human-originated, and are therefore
604−symbolic, but there are exceptions. For example, blogposts may
612+symbolic, but there are exceptions. For example, blogposts may
605613want to be named by date.
606614
607615However, it is slightly retarded for the fundamental state of our
608616arches to store the _rendered_ string (`++knot`), not the source
609−data (`++coin`). It leads to all kinds of inefficiencies where
610−we're doing way more parsing than we want to. (Parsing also
611−should be jet-propelled, but is not.) It also makes it way too
617+data (`++coin`). It leads to all kinds of inefficiencies where
618+we're doing way more parsing than we want to. (Parsing also
619+should be jet-propelled, but is not.) It also makes it way too
612620easy to sneak in bad data and make it `%clay`'s problem.
613621
614622So, logically, we'd be shifting from
@@ -623,7 +631,7 @@ So, logically, we'd be shifting from
623631
624632Also, a vanilla `++map` might not be quite the right data
625633structure at this point -- we probably want a custom sorted tree,
626−which separates auras and imposes a custom order on each. This
634+which separates auras and imposes a custom order on each. This
627635would let us serve simple range queries quite efficiently, so
628636long as they match the standard aura order (eg, alphabetical
629637order for symbols).
@@ -631,23 +639,23 @@ order for symbols).
631639### Change #5: dome semantics by reference
632640
633641While the principle that a dome contains its own mark source is
634−elegant, it's a pain in the ass to implement properly. `%clay`
642+elegant, it's a pain in the ass to implement properly. `%clay`
635643certainly does not do it justice at present.
636644
637645The complexity is in checking every change to see whether it
638646includes a change to the mark source (which is very unlikely),
639647or a dependency of the mark source (slightly less unlikely).
640648And if so, do we have to update all the leaves in the dome?
641−This is also very unlikely. And yet, it happens.
649+This is also very unlikely. And yet, it happens.
642650
643651Instead, we could simply designate, by beak (ship, desk, and
644−case) a dome which contains the canonical mark source. This
652+case) a dome which contains the canonical mark source. This
645653would rarely change -- you could upgrade a desk to a new mark
646654dome, but it would be a manual operation and probably a rare one.
647655
648656The common case would just be that everyone uses the current
649657version of the `%base` desk when a desk is created, and never
650−changes it. In practice, this means not every two-byte edit
658+changes it. In practice, this means not every two-byte edit
651659needs to get involved in detecting and/or applying schema
652660evolution, a welcome simplification.
653661
@@ -659,14 +667,14 @@ design is one mark per message structure, the mark namespace gets
659667overloaded fast.
660668
661669So we added an app hierarchy within the mark namespace, using hep
662−`-` as a divider. The `:talk` app expects messages with the mark
670+`-` as a divider. The `:talk` app expects messages with the mark
663671`talk-message`, whose sources is in `/===/mar/talk/message`.
664672
665673This helps, but it doesn't solve another two problems: trivial
666674nouns and containers.
667675
668676It's retarded to have to add a source file to send a `@ud` or a
669−`*`. An aura (see above under `coin`) should be a trivial mark.
677+`*`. An aura (see above under `coin`) should be a trivial mark.
670678
671679Also, we need to be able to construct standard containers of
672680marks: lists, maps, sets, and tuples (or, as a database person
@@ -678,32 +686,32 @@ Another kind of complex mark that needs to exist: conflict,
678686difference, and invalid marks.
679687
680688Conflict data and difference data are two types of noun that are
681−generated in normal RCS operation. If we have the concept of an
689+generated in normal RCS operation. If we have the concept of an
682690`%html` noun, we also need the concept of an invertible change
683691from one `%html` file to another; a merge conflict between two
684692`%html` files; and even a conflict between two marks.
685693
686694There are two kinds of changes to a dome: manual and automatic.
687695Manual changes are initiated by the user and should crash, unless
688−manually overridden. Automatic changes should not crash, ever,
689−and should not lose data. If they create problems, which they
696+manually overridden. Automatic changes should not crash, ever,
697+and should not lose data. If they create problems, which they
690698shouldn't, these problems will need to be cleaned up manually.
691699
692700This is a complex space to navigate, but it requires a conflicted
693701merge to be able to resolve to a valid dome containing conflict
694−leaves, not crash. The present `%clay` goes halfway there and
702+leaves, not crash. The present `%clay` goes halfway there and
695703resolves to a scratch desk, but we can go farther: the whole `git`
696−dumpster-fire of "merge state" deserves destruction. One change
704+dumpster-fire of "merge state" deserves destruction. One change
697705is a merge that produces conflicts, the next is an edit that
698706resolves them.
699707
700−Similarly, all sorts of errors can produce invalid leaves. One
708+Similarly, all sorts of errors can produce invalid leaves. One
701709particularly rare case which is nonetheless important is schema
702710evolution: what happens to files in a mark, when the mark source
703−code changes. They all need to be re-normalized. And this
704−operation can fail, producing an invalid value. (We don't do any
711+code changes. They all need to be re-normalized. And this
712+operation can fail, producing an invalid value. (We don't do any
705713of this now, BTW, although `%ford` has the dependency mechanism to
706−make it possible.) Again, in some cases this should crash the
714+make it possible.) Again, in some cases this should crash the
707715update to the mark source; in other cases, there is no local way
708716to handle the crash.
709717
@@ -714,46 +722,46 @@ into using the bad content.
714722
715723The bottom line: a mark needs to be split into two models:
716724`++mark`, the original symbolic mark with its source engine, and
717−something like `++logo`, the meta-mark or mark container. The
725+something like `++logo`, the meta-mark or mark container. The
718726mark is the building block; the logo is the building.
719727
720728### Change #8: filesystem hierarchy standard
721729
722730While all filesystems remain forks of one ur-filesystem, the
723−standard tends to enforce itself. But not define itself.
731+standard tends to enforce itself. But not define itself.
724732
725−I've always hated the current `/mar` for marks, for instance. I
726−prefer `/mac`. But others hate this. And so on. Also, by
733+I've always hated the current `/mar` for marks, for instance. I
734+prefer `/mac`. But others hate this. And so on. Also, by
727735necessity, our nascent "FHS" is heavily biased toward system
728−data, away from user data. This is retarded.
736+data, away from user data. This is retarded.
729737
730−Please focus all bikeshedding energy on this problem! Better to
738+Please focus all bikeshedding energy on this problem! Better to
731739bikeshed about what color to paint the bikeshed, then what shape
732740the bike gears should be.
733741
734742### Change #9: security and access control
735743
736−Oh, yeah, that! They always say not to design security in at the
737−last minute. Unfortunately, this is exactly what we did. There
738−is no security at all in present `%clay`. (There is also no push
744+Oh, yeah, that! They always say not to design security in at the
745+last minute. Unfortunately, this is exactly what we did. There
746+is no security at all in present `%clay`. (There is also no push
739747operation in `%clay`, so it's not as mad as you think -- anyone
740748can read anything, no one can write anything, which is fine in a
741749young network where there are no secrets.)
742750
743751Fortunately, we don't have anything super fancy planned in this
744−area, just basic access-control lists. These obviously need to
752+area, just basic access-control lists. These obviously need to
745753be designed along with the filesystem-hierarchy standard (see
746754elsewhere).
747755
748−Is it even possible to do ACLs properly in `%clay`? At this
749−point, yes. There is an event-oriented security model in Arvo,
756+Is it even possible to do ACLs properly in `%clay`? At this
757+point, yes. There is an event-oriented security model in Arvo,
750758though it isn't really hooked up to anything at the moment.
751−There is also a privilege badge system in `%jael`. The
759+There is also a privilege badge system in `%jael`. The
752760combination of these two features is quite sufficient for a
753761decent ACL model.
754762
755763The idea of event-based security is that every event passed to
756−`%clay` has a security `++mask`. The mask tells us:
764+`%clay` has a security `++mask`. The mask tells us:
757765
758766- the set of ships that has _contaminated_ this event
759767- the set of ships that has _endorsed_ this event
@@ -761,20 +769,20 @@ The idea of event-based security is that every event passed to
761769
762770To simplify this system a little and maybe explain it, the
763771security question in any event is: (a) who has _caused_ this
764−event? (b) who may receive its _effects_? (These sets never
772+event? (b) who may receive its _effects_? (These sets never
765773include your urbit itself, of course, only neighbors.)
766774
767775For example, if we're doing something because ship X sent us a
768776packet, and we're going to leak the results of this something
769777back in a response to ship X, once `%clay` knows what our ship
770778thinks of its neighbor X, we should have a good picture of what
771−effects this event should be allowed to cause. If `%jael` knows
779+effects this event should be allowed to cause. If `%jael` knows
772780that X is our friend, `%clay` can know that X can see our cat
773781picture.
774782
775783Causality is separated out into contamination (data from ship X
776784got into the event) and endorsement (ship X signed the whole
777−event). (This may be overdesign.) All these sets can be defined
785+event). (This may be overdesign.) All these sets can be defined
778786as the empty set (eg, releasing to the empty set means no output)
779787or the universal set (input from the universal set could be from
780788anyone, ie, is unauthenticated).
@@ -783,11 +791,11 @@ anyone, ie, is unauthenticated).
783791
784792When you read a leaf from a `%clay` dome, `%clay` is happy to
785793perform any necessary mark conversions (the actual work is done
786−by the `%ford` build vane). You can also just ask for the leaf
794+by the `%ford` build vane). You can also just ask for the leaf
787795in whatever mark `%clay` has it.
788796
789797It's confusing to think of this mark extension as part of the
790−path. You shouldn't say `+cat /===/app/talk/hoon`. You should say
798+path. You shouldn't say `+cat /===/app/talk/hoon`. You should say
791799`+cat /===/app/talk`, or maybe `+cat %hoon /===/app/talk`.
792800
793801A mark-extended namespace may be added to `%ford`, to compensate.
@@ -796,7 +804,7 @@ But type-converting reads are the rare case, not the common case.
796804### Change #11: real queries and indexes
797805
798806A filesystem should postpone being a database until it can no
799−longer do so. So, please don't think about this problem yet.
807+longer do so. So, please don't think about this problem yet.
800808I'm joking, of course!
801809
802810### Change #12: meta-query
@@ -805,24 +813,24 @@ Make it possible to query the list of desks in a ship :-)
805813
806814### Change #13: completely redesign Unix mounting
807815
808−This is actually a series of changes, described below. But
816+This is actually a series of changes, described below. But
809817broadly...
810818
811819One of Urbit's experiments is live two-way synchronization
812820between the Unix filesystem and the revision-control store.
813−Dropbox style, as it were. I think this experiment is basically
814−a failure. It may be right for cat pictures, but I don't think
821+Dropbox style, as it were. I think this experiment is basically
822+a failure. It may be right for cat pictures, but I don't think
815823it's right for Urbit.
816824
817825The fundamental problem is that there are all sorts of weird,
818−flaky ways applications use a mutable filesystem. It's not at
826+flaky ways applications use a mutable filesystem. It's not at
819827all unusual for an editor to change a file by deleting the old
820828file, then writing the new one in its place.
821829
822−You really don't want this to produce a delete change. A simple
830+You really don't want this to produce a delete change. A simple
823831cooldown timer for the sync helps a lot with this, but causes
824−other weirdnesses. In theory, a sync can even cause another
825−change to the same version of the same desk. And so on.
832+other weirdnesses. In theory, a sync can even cause another
833+change to the same version of the same desk. And so on.
826834
827835The thing we fear is just too funky to work (there is some
828836strange "reflux bug" still affecting it) is any sort of _two-way_
@@ -833,8 +841,8 @@ cold storage, is that there's no useful workflow that actually
833841uses it -- at least, when editing Urbit code.
834842
835843It's actually really nice to have files on a disk that nothing
836−will automatically modify. If you're using `git` and `%clay` in
837−parallel, it's nice to mount your `git` working directory. The
844+will automatically modify. If you're using `git` and `%clay` in
845+parallel, it's nice to mount your `git` working directory. The
838846original Urbit workflow was to work out of a mounted directory,
839847but the two-way updates were just too weird and scary, so now we
840848tend to edit elsewhere and copy in -- which is lame.
@@ -847,25 +855,25 @@ FUSE (in a somewhat experimental state), and filesystem watching
847855
848856The FUSE interface is a good one, though it needs more work.
849857Moreover, it exports the whole Urbit namespace to Unix, not just
850−`%clay`. For reading Urbit data, FUSE is basically ideal.
858+`%clay`. For reading Urbit data, FUSE is basically ideal.
851859
852−The obvious impulse is to make a read-write FUSE. But I don't
860+The obvious impulse is to make a read-write FUSE. But I don't
853861think FUSE should be used for recording changes to an RCS, even
854862though the FUSE API sends you changes as events.
855863
856−There are three reasons for this. One, the actual filesystem
864+There are three reasons for this. One, the actual filesystem
857865writes that Unix apps make are typically of low quality and
858−confusing. Some crack-brained editor will find it completely
866+confusing. Some crack-brained editor will find it completely
859867normal and unremarkable to modify a byte in a file, by deleting
860−the file and rewriting it from scratch. You can use a cooldown
868+the file and rewriting it from scratch. You can use a cooldown
861869timer to wait for this nonsense to converge, but leaving data in
862870heuristic limbo seems like a very bad idea in FUSE-land.
863871
864872The second is that people derive a lot of comfort from seeing
865−actual files on an actual disk. The idea that your datas are
873+actual files on an actual disk. The idea that your datas are
866874stuck in Urbit is slightly terrifying as a user experience.
867−We claim to be reliable. It's actually hard to imagine datas that
868−couldn't be recovered from either snapshot or log. But how can
875+We claim to be reliable. It's actually hard to imagine datas that
876+couldn't be recovered from either snapshot or log. But how can
869877you possibly believe us on this?
870878
871879So I think it's correct to lean all the weight of reading data
@@ -880,35 +888,35 @@ But this is a separate discussion.
880888
881889### Change #15: one-way, nonced working desk
882890
883−A one-way mount is a much less crazy thing. It works the same
891+A one-way mount is a much less crazy thing. It works the same
884892way our two-way mount works today, but with only one writer: the
885893user, via Unix.
886894
887895Since a mount is one-way, it needs to have an absolute write lock
888−on the desk it's writing to. We cannot have a `%home` desk which
896+on the desk it's writing to. We cannot have a `%home` desk which
889897we are also editing with filesystem watching.
890898
891899Instead, we have a temporary and anonymous state that's connected
892−itself with your `%home` desk. Desk names have to be valid
893−symbols, but we can use `hep`, `-`. A reasonable choice is to
900+itself with your `%home` desk. Desk names have to be valid
901+symbols, but we can use `hep`, `-`. A reasonable choice is to
894902pick a random 32-bit name for each mount instance -- so the desk
895−name is not `%home`, but `%home-bosdut-sigder`. (Randomizing
903+name is not `%home`, but `%home-bosdut-sigder`. (Randomizing
896904instead of always using a fixed suffix, say, `%home-work`,
897905just makes all kinds of screwups less likely.)
898906
899907You create a mount simply by mkdiring the desk name in your pier.
900908Just the base name, Urbit will come up with the `%bosdut-sigder`
901−part and put it in a `.desk` file at the desk root. To forget
902−the mount, just delete this file. This avoids the ugly confusion
909+part and put it in a `.desk` file at the desk root. To forget
910+the mount, just delete this file. This avoids the ugly confusion
903911between unmounting a mount, and deleting all the data in it.
904912
905913(Urbit will keep a set of the mount desks it's used and
906−forgotten, not reuse them, and block all requests into them. By
914+forgotten, not reuse them, and block all requests into them. By
907915blocking, we can delete without losing referential transparency.
908916Yes, Virginia, Urbit can forget!)
909917
910918Note that the auto-reload behavior of applications is extremely
911−useful when testing. When testing, you can either switch the
919+useful when testing. When testing, you can either switch the
912920active application desk from `%home` to `%home-bosdut-sigder`,
913921or activate autosync from the latter to the former.
914922
@@ -931,7 +939,7 @@ Otherwise, you're basically, like, reinventing Perforce.
931939
932940More concretely, the Urbit experience today requires `:hood`
933941pokes (Urbit shell commands, basically) for all nontrivial RCS
934−operations, from mounting to merging. Again, there are lots of
942+operations, from mounting to merging. Again, there are lots of
935943situations in which you can write into a mounted filesystem (the
936944protocol for Unix communication could be easily adapted to
937945communicate with a browser-based editor, for instance), but you
@@ -941,20 +949,20 @@ Even commands like mounting and merging are easily communicated,
941949through filesystem writes that any Unix tool (including many
942950macro editors).
943951
944−This is very easy for a filesystem watcher. One such protocol:
952+This is very easy for a filesystem watcher. One such protocol:
945953if you want to send the daemon a command relative to some
946−directory, write a `.ur-command` metafile in that directory. The
954+directory, write a `.ur-command` metafile in that directory. The
947955daemon accepts the command by deleting it.
948956
949957The user experience is that you configure your mounts by simply
950−making the directories and configuring dotfiles in them. Unlike
958+making the directories and configuring dotfiles in them. Unlike
951959in the present system, Urbit will never modify your datas
952960silently in the background.
953961
954962Here's what this looks like:
955963
956−Suppose you want to edit some data in the `%home` desk. Your
957−data is in `/=home=/mar/talk`. You also want a read-only view
964+Suppose you want to edit some data in the `%home` desk. Your
965+data is in `/=home=/mar/talk`. You also want a read-only view
958966of `/=home=/arvo`.
959967
960968From Unix, you do something like this:
@@ -967,18 +975,18 @@ From Unix, you do something like this:
967975 $ touch home/arvo/.read
968976```
969977
970−Want to commit your changes? That means merging the working
971−desk, `%home-bosdut-sigder`, into the `%home`. Put a commit message
972−into `home/.commit` and it'll happen. If there's an error, it
978+Want to commit your changes? That means merging the working
979+desk, `%home-bosdut-sigder`, into the `%home`. Put a commit message
980+into `home/.commit` and it'll happen. If there's an error, it
973981will appear as `.error`.
974982
975983The mount has a .version file which simply contains the change
976−number (of the working desk) which any edits are relative to. If
984+number (of the working desk) which any edits are relative to. If
977985the urbit daemon is running and watching, edits will be absorbed
978986quickly (probably after a heuristic timeout), and the version
979−incremented. But we always are confident that the edits are
987+incremented. But we always are confident that the edits are
980988against the latest version, or at least we can use this as a
981−sanity check. All kinds of insane things can happen to get a
989+sanity check. All kinds of insane things can happen to get a
982990filesystem out of consistency.
983991
984992This design obviously needs a little fleshing out, but I'm quite
@@ -987,12 +995,12 @@ from Unix.
987995
988996## In conclusion
989997
990−There's obviously a lot of fun here! There are two things you
998+There's obviously a lot of fun here! There are two things you
991999can do right now.
9921000
9931001One, talk about `%clay` on [this forum
9941002thread](http://urbit.org/~~/fora/posts/~2016.10.25..17.37.47..6937~/) or
995−on [:talk](http://urbit.org/stream). Both hosted on Urbit
1003+on [:talk](http://urbit.org/stream). Both hosted on Urbit
9961004itself!
9971005
9981006Two, poke around the