1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
|
Document: WM-042 P. Webb
Category: Operating Systems 2019.12.02
A personal API
Abstract
You are the operating system
Body
1. Inspiration
Shortly after revisiting[1] my thoughts about the future of
operating systems I was pestered with a recurring thought; what if
one had a personal API sitting on a $5/month VPS (virtual private
server) and interfaced with it for nearly everything they do
online? It wasn't far-fetched for me to then imagine this personal
API becoming more of a personal OS in a way (or maybe stylized,
personalOS). With that in mind I realized that when thinking about
updating my personal web page[2] I didn't want to update the
content manually. Rather, I'd want to update an API and have my
site(s) automatically update. *Oh yeah*, why NOT have multiple
sites update?
A personal API would make it easy for people to decentralize parts
of their digital lives. There are companies built upon the concept
of decentralization and that's promising. A rising tide raises all
ships, as they say, and I can see a day where non-technical people
are spinning up servers as much as they're migrating data to a new
smartphone (well, if they did all this via a nice UI).
2. Use Cases
Here's a use case: I'm signed up to several social networks,
forums, and I have a couple websites. I'm a digital tastemaker and
pride myself on being current and relevant. However, updating my
account bio across these platforms is a 20-minute task that I
dread doing. Never fear for my API has a biography module!
Realistically, all the platforms I frequent won't have an API I
can connect my own to, but a handful will. My personal API sends
out update requests to the platforms it is connected to and
updates my bio section when I publish the relevant changes to
my API.
That example may be a little contrived so let's try another one:
I'm a developer who prefers to self-host things whenever possible.
I am also quite lazy and live by DRY (don't repeat yourself) and
KISS (keep it simple, silly) principles to the extent I can
without much difficulty. I'm also fatigued and distrustful of the
capability of online services in general to keep my data safe. If
I trust nothing else I at least trust myself and a code base I
can peruse.
3. Version Alpha
*Please keep in mind that this is still a super early idea and the
purpose of this post is to get the particulars fleshed out.*
What follows is my proposed folder structure of the personal API.
Modules come in two types: purpose and utility. Utility modules
allow the purpose modules to do their thing, like
retrieving/modifying/storing data (on DigitalOcean's Spaces,
Amazon's S3, or your own storage location). Ideally you should
never have to touch those modules unless you wanted to improve
upon it or fork your own version. Purpose modules can be anything
you think of and the examples I listed could be your personal
Medium, Pinboard/Pinterest, Things, Spotify, Flickr/Instagram,
Last.fm, and so on.
```text
api
├─ …
├─ modules
│ ├─ purpose
│ │ ├─ blog
│ │ ├─ bookmarks
│ │ ├─ music
│ │ ├─ photos
│ │ ├─ portfolio
│ │ ├─ reminders
│ │ ├─ scrobbler
│ │ └─ status
│ └─ utility
│ ├─ database
│ └─ storage
└─ …
```
The purpose functions would be able to call the utility functions
to, well, function. `webb.page` will be the guinea pig for the
initial release of personalOS. Once it works there I will update
parts of `inc.sh` and `dsgn.agency` to take advantage of it
as well.
4. Why Bother?
You, dear reader, are probably not terribly interested in
self-hosting and dev ops. Even developers/engineers who know HOW
to do it aren't interested. Some of the common reasons against
self-hosting are fear (of the work involved), apathy ("my data
isn't super important anyway"), time (the only valid reason IMHO),
or a combination of the three. While I understand I also think *it
is incredibly important to claim your space on the Internet*.
We're all digital sharecroppers[3] so why not tend to our own
space on the 'Net?
It's a helluva lot easier to monetize your content when it lives
on your platform, or personalOS in this case. Monetization doesn't
have to be your focus though. How about this: something that
irritates me about Apple's Music app and iOS 13 is that I have no
idea how to get my music that used to live in iTunes, onto my
phone anymore. Once I build a music module for personalOS I can
upload my music library to my own server and stream it on the go.
The major streaming services are great and all but my music
collection has rips from video game soundtracks, anime, and
otherwise defunct groups with no content on these streaming
services. Why would I give that up? Why *should* I?!
5. Business Opportunity?
Y'know what would be nice? Having an easy way to share any files
that you're hosting. Although, you may not want to expose your
personal API server to the public. For this purpose, you could use
a third-party site like `personal.sh` to generate shareable links
(probably via `personal.sh/are`, I'm a fan of domain hacking[4]).
Because the Internet is open to all, there will inevitably be
disturbing/disgusting/morally-bankrupt content shared through
`personal.sh`. If/when that happens, the offending user will have
their IP address blacklisted and put on display at
`personal.sh/it`. It's no secret that anonymity gives some people
courage to behave in a way that's unbecoming.
6. Prior Art
There's a company named Urbit whose mission is quite similar:
> An Urbit is a networked personal server that runs your apps and
> stores your data, serves as your permanent online identity and
> puts you back in control of your digital life under a
> single login.
>
> When you’re ready.
> — urbit live[5]
Only problem is, their official site makes their mission quite
confusing[7] to many[8]. I don't know about you but if I'm
entrusting my data to something I run, I want to have more than a
*general idea* of how it works.
*EDIT (2019.12.05):* This post has generated lively discussion on
HackerNews[9] and as a result we now have more examples of
prior art!
- cjdns[10]
- Dogsheep[11], by Simon Willisons
- Islet[12], by Carl Hewitt
- libp2p[13]
- me-api[14], by Daniel Fang
- my[15], by Dmitrii Gerasimov
- Perkeep[16]
- SOLID (Social Linked Data)[17], by Tim Berners-Lee
- Syte2[18] / Syte2 for Zeit Now v2[19], by Rodrigo Neri and
Jake Peterson
- Unwalled Garden[20]
- Yggdrasil Network[21]
FIN
Elsewhere, there exists portions of the personalOS concept of
various development maturation levels. To me, this is proof of the
worthwhile nature of this idea. Will this likely end up sitting in
a repo for years after months of initial work? Probably. Hopefully
not…we'll see how it goes. 🕸
References
[1] </WM-039>
[2] <https://webb.page>
[3] <https://info.ideagrove.com/knowledgebase/digital-marketing/what-is-digital-sharecropping-and-why-is-it-a-bad-thing>
[4] <https://en.wikipedia.org/wiki/Domain_hack>
[5] <https://twitter.com/urbitlive/status/1192262246384230400>
[6] <https://news.ycombinator.com/item?id=21672481>
[7] <https://news.ycombinator.com/item?id=18908051>
[8] <https://news.ycombinator.com/item?id=6438320>
[9] <https://news.ycombinator.com/item?id=21684949>
[10] <https://github.com/cjdelisle/cjdns>
[11] <https://dogsheep.github.io>
[12] <https://www.infoq.com/presentations/strong-types-actor-iot>
[13] <https://libp2p.io>
[14] <https://github.com/danfang/me-api>
[15] <https://github.com/karlicoss/my>
[16] <https://perkeep.org>
[17] <https://solid.mit.edu>
[18] <https://github.com/rigoneri/Syte2>
[19] <https://github.com/jake-101/Syte2-4-Zeit2.0>
[20] <https://unwalled.garden>
[21] <https://github.com/yggdrasil-network/yggdrasil-go>
|
