11 |

12 | 13 |

Personal

14 |

15 | Paolo Fragomeni, Software Engineer

16 | 17 |

Summary

18 |

19 | I left MIT in 2010 to co-found my first startup. Since then I have founded, 20 | funded, and sold multiple companies. I'm an engineer, founder, and startup advisor. 21 | 22 |

Expertise

23 |

24 | Computer Science Research. Software Engineering: programming design, 25 | engineering, concurrent and distributed systems programming. 26 | 27 |

Experience

28 |

Founder, Engineer, CEO — Socket Supply Co.

29 |

30 | Dec 2021 - 31 |

32 | Socket Supply Co. makes peer-to-peer practical for the average web developer, 33 | allowing them to build cross-platform apps that connect directly to each 34 | other, completely obsoleting the need for costly and complex servers. 35 | 36 |

Technologist — Passport Capital

37 |

38 | Oct 2018 - Jan 2020 39 |

40 | Passport Capital is a San Francisco based hedge fund. My role was to provide 41 | technical venture diligence. 42 | 43 |

CTO, Cofounder — Voltra Co.

44 |

45 | Jan 2016 - Aug 2018 46 |

47 | Voltra Co. was a decentralized music store and player with a focus on 48 | music ownership. Electron, JavaScript, Node.js, C++, CSS3, HTML5, Stylus, Webpack, 49 | React, React-Native, and Amazon Web Services. Voltra joined Passport Captail 50 | in October 2018. 51 | 52 |

Organizer — Data Tera Nemo

53 |

54 | Feb 2015 - May 2019 55 |

56 | A 2 day computer science conference with a focus on distributed systems. In 2015 57 | and '19 this conference brought together the engineers who created WebTorrent, Dat 58 | (Now HyperProtocol), libp2p (before and after Protocol Labs was founded), 59 | Scuttlebutt, Beaker, and many contributors to the most significant projects in the 60 | space. 61 | 62 |

VP of Engineering — Now Secure

63 |

64 | Nov 2014 - Jan 2016 65 |

Engineer, CTO, Cofounder — Here Is How

71 |

72 | Nov 2012 - Jan 2014 73 |

74 | A CMS for technical writing, a web based interface similar to Medium.com capable of safely 75 | executing arbitary code for any platform. Docker, JavaScript, Node.js, Websockets, C, C++, 76 | HTML5, CSS3. 77 | 78 |

Engineer, CTO, Cofounder — Nodejitsu

79 |

80 | Sep 2010 - Dec 2012 81 |

82 | Co-Founder, Chief Technology Officer. Lots of R&D. Conceptualized and 83 | implemented products that simplify and manage application deployments 84 | for the node.js platform. 85 |

82 |

83 |

a little class for combining behavior, layout and style

84 |

2018-7-30

85 |

I like some of what React offers. I also think it's bloated, over-hyped, 86 | over-engineered and it sees the web as a compile target rather than a 87 | development platform.

88 |

I like most of what Web Components offer, they're a part of the 89 | web platform. They offer real encapsulation — for css. 90 | And like most web APIs designed by consensus, they're awkward.

91 |

92 |

Tonic is about 250 lines of code. It borrows goals and ideas from React but is 93 | built on native Web Components. It works in all browsers. It's stable. It's the 94 | minimum of what is needed to organize application code, the flow of data and 95 | accommodate component based architecture.

96 |

You can find the core library here and a collection of components here 97 | on Github.

98 |

2019-7-3 Update

99 |

Tonic is about a year old. To celebrate a year without any new features, let's 100 | add a new feature...

101 |

Your render function can now be async or an async generator. This 102 | provides a declaritive way to express the intent of your render method. For 103 | example...

104 |

class SomeComponent extends Tonic {
105 |   async * render () {
106 | 
107 |     yield loadingState()
108 | 
109 |     return await getStuff()
110 |   }
111 | }

112 | 113 |

114 | 117 |

118 |

119 | 120 |

121 |

What are javascript then-ables

122 |

async / await improves program flow and reduces the number of callbacks in 123 | your code. The await keyword can be used to pause the current code path and 124 | wait for an async function to return a value without blocking the event loop.

125 |

async function main () {
126 |   console.log('started')
127 |   await sleep(100)
128 |   console.log('finished')
129 | }
130 | 
131 | main()

132 |

An implementation for the sleep function might look like this...

133 |

const sleep = n => new Promise(r => setTimeout(r, n))

134 |

However, as this example demonstrates, the code in the promise starts executing 135 | immediately. Promises are eager (as opposed to being lazy), and sometimes we 136 | want them to be lazy. Here is a detailed explaination of why an eager 137 | promise may not be what you want.

138 |

A then-able is lazy. It has no constructor. It's any function, object or class 139 | that implements a then method.

140 |

Await-able Classes

141 |

To create an async class, implement a then method on it!

142 |

class Foo {
143 |   then (resolve, reject) {
144 |     resolve(42)
145 |   }
146 | }
147 | 
148 | async function main () {
149 |   const answer = await new Foo()
150 |   // answer === 42
151 | }
152 | main()

153 |

Await-able Objects

154 |

You can do the same thing with an object. You can name the callback 155 | functions whatever you want. Also, you aren't required to use or care 156 | about the rejection callback.

157 |

const Foo = {
158 |   then (resolve) {
159 |     setTimeout(() => resolve(42), 1024)
160 |   }
161 | }
162 | 
163 | async function main () {
164 |   const answer = await Foo
165 |   // answer === 42
166 | }
167 | main()

168 |

Await-able object factories

169 |

const Foo = num => ({
170 |   then (resolve) {
171 |     setTimeout(() => resolve(num), 1024)
172 |   }
173 | })
174 | 
175 | async function main () {
176 |   const answer = await Foo(42)
177 | }
178 | main()

179 |

Async then-ables

180 |

Object and class methods can use the async keyword, just like functions.

181 |

const Foo = {
182 |   async then (resolve) {
183 |     resolve(await request('https://foo.com'))
184 |   }
185 | }

186 |

Destructuring assignments provide a way to return multiple values...

187 |

class Foo {
188 |   then (resolve) {
189 |     request('https://foo.com', (err, res) => resolve({ err, res }))
190 |   }
191 | }
192 | 
193 | async function main () {
194 |   const { err, res } = await new Foo
195 | 
196 |   // More than one err? Const is block-scoped!
197 |   {
198 |     const { err, res } = await new Foo
199 |   }
200 | 
201 |   // Destructured values can also be aliased.
202 |   const { err: namedError, res: namedResponse } = await new Foo
203 | }
204 | main()

205 | 206 |

207 | 210 |

211 |

212 |

213 |

preamble

214 |

synopsis

215 |

This is a collection of notes that explore peer-to-peer topics.

216 |

description

217 |

Rhis collection focuses on the following topics...

218 |

Connectivity
Replication
State
Consensus
Consistency
Security

226 |

These notes are not complete and don't advocate any particular approachs. 227 | They are related to my work on dat-cxx.

228 | 229 |

230 | 233 |

234 |

235 | 236 |

237 |

illustrated lamport timestamp

238 |

problem

239 |

With the client-server model, you can easily determine the order of 240 | events in a system because they are all maintained by a single source. 241 | This is critical in, for example a chat application.

242 |

But with the distributed model, how do we know if an event happened 243 | before another? How can we thread together datasets from different 244 | sources in the correct order?

245 |

solution

246 |

A Lamport Timestamp is one solution to determine the order of events 247 | in a distributed system. Although it may not solve all problems in this 248 | problem space, it is a useful primitive that we will explore.

249 |

Clocks vs. Logical Clocks

250 |

Why don't we use regular time stamps? Most clocks count time at different rates 251 | and experience failures that require resynchronization. This means they are 252 | reliably unreliable for determining the order of events.

253 |

Lamport Timestamps use a Logical Clock to keep track of the order of events 254 | on each node. A logical clock is not a clock that keeps track of the time, it's 255 | a monotonically increasing counter. So, when a node in a network receives a message, it 256 | re-synchronizes its counter (its clock) with the node that sent the message.

257 |

Example

258 |

Node A increments its clock before each event that hapens. An event is 259 | something meaningful, like when it creates some data. When node A 260 | eventually sends its payload over the network, it will include the current 261 | value of its clock.

262 |

let clock = 0
263 | 
264 | //
265 | // A thread-safe, monotonically increasing function.
266 | //
267 | function createTimeStamp () {
268 |   clock += 1
269 |   return clock 
270 | }
271 | 
272 | function doSomething (data) {
273 |   //
274 |   // Do something with some data.
275 |   //
276 |   return {
277 |     data,
278 |     clock: createTimeStamp()
279 |   }
280 | }
281 | 
282 | //
283 | // Eventually send the data to the network.
284 | //
285 | sendToNetworkQueue(doSomething({ ... }))

286 |

When node B receives a message from node A, it will decide how to set 287 | its own clock. If the clock value in the message is greater than its own 288 | value, it will use the value in the message. Otherwise it will use its own 289 | value. In either case, it will also increment its own clock by 1.

290 |

let clock = 0
291 | 
292 | //
293 | // Eventually receive some data from the network.
294 | //
295 | receiveFromNetworkQueue (message) {
296 |   clock = Math.max(message.clock, clock) + 1
297 | }

298 |

Here we semi-randomly fail to always tell the other node about the event that 299 | happened, illustrating what happens when a node is eventually synchronized.

300 |

301 |

302 |

303 |

304 | 305 |

306 |

Fire event on node A

307 |

Fire event on node B

308 |

309 | 310 |

This may not be the correct primitive for all your use cases. For example, 311 | Lamport Timestamps don't express causality, meaning, the reason why one 312 | event happened before another isn't in scope of this soluton, but that is 313 | something that can be achieved using a Vector Clock.

314 | 405 | 406 |
407 | 408 |

Special thanks to Fedor Indutny and Feross Aboukhadijeh for reviewing 409 | this post. ♥

410 | 411 |

412 | 415 |

416 |

417 | 418 |

419 |

vector clocks

420 |

In the previous post, I wrote about how Lamport Timestamps (aka Logical 421 | Clocks) can help determine the order of events in a distributed system.

422 |

problem

423 |

Logical clocks only offer "Partial Ordering", because they can tell us the 424 | order of a single event, but not the total ordering of events or why 425 | a system arrived at its state.

426 |

solutions

427 |

Vector Clocks build on the idea of Logical Clocks to help track 428 | causality in a distributed system.

429 |

Here is an example vector clock in a network where there are three 430 | participating nodes...

431 |

{ alice: 0, bob: 1, carol: 0 }

432 |

To set up a node we will give it an id and an in memory object to 433 | store some data.

434 |

const myId = 'alice'
435 | const data = {}

436 |

sending messages

437 |

When a node writes some data, it increments its own logical clock in 438 | the vector and includes it as a property of a message that it will 439 | attempt to send. We also add the value as a property of the message.

440 |

function write (key, value) {
441 |   if (!data[key]) {
442 |     data[key] = { clock: { [myId]: 0 } }
443 |   }
444 | 
445 |   data[key].clock[myId] += 1 
446 |   data[key].value = [value]
447 | 
448 |   send(key, data[key])
449 | }

450 |

451 |
In this case we made the value property an array. This is because we 452 | must anticipate the possibility of concurrent messages — that is, 453 | a message was received where two nodes have a logical clock with the same 454 | count.
455 |
In this case we can push the new value onto the array and allow the the 456 | conflict to be resolved somehow (we'll discuss this more later).
457 |

458 |

receiving messages

459 |

When a node receives a message it increments its own Logical Clock in 460 | its local copy of the vector.

461 |

Then for each node in the message's vector, it compares the local 462 | clock count (if there is one) to the clock count in the received 463 | message, taking the max of the numbers.

464 |

const max = arr => Math.max.apply(null, Object.values(arr))
465 | 
466 | function receive (message) {
467 |   const key = message.key
468 | 
469 |   //
470 |   // If this is new data, there is no need to compare anything.
471 |   // we can store it locally and return early from the function.
472 |   //
473 |   if (!data[key]) {
474 |     data[key] = message
475 |     data.clock[myId] = max(message.clock) + 1
476 |     return
477 |   }
478 | 
479 |   //
480 |   // We have received the message, update our clock
481 |   //
482 |   data[key].clock[myId] += 1
483 | 
484 |   const localClock = data[key].clock
485 |   const messageClock = message.clock
486 | 
487 |   //
488 |   // For each node in the vector of the message
489 |   //
490 |   for (const id in Object.keys(messageClock)) {
491 |     const a = localClock[id] || 0
492 |     const b = messageClock[id]
493 | 
494 |     const isConcurrent = a === b
495 | 
496 |     if (isConcurrent) {
497 |       data[key].conflict = true
498 |       data[key].value.push(message.value)
499 |       continue
500 |     }
501 | 
502 |     const happenedBefore = a < b
503 | 
504 |     if (happenedBefore) {
505 |       data[key].value = [message.value]
506 |     }
507 | 
508 |     localClock[id] = Math.max(a, b)
509 |   }
510 | }

511 |

handling concurrent messages

512 |

Two messages that are received at the same time and have the same logical clock 513 | count are "concurrent".

514 |

To understand what to do with this type of data, we need to create a resolution 515 | function. This function may be the only way to determine what data is either a 516 | descendant or which data comes-before.

517 |

Reject the data and send it back to the clients asking for it to be resolved. 519 | This might mean asking them to manually merge or discard some of the data.
520 |
Last-Writer-Wins uses time-based timestamps. If you consider clock-drift 522 | (mentioned in the first post), there is a high probability of losing data with 523 | this strategy.
524 |

526 |

research timeline

527 |

When discussing Vector Clocks we should consider some other closely related 528 | research...

529 |

530 | 531 | Lamport Timestamp 532 | 1978 533 | 534 | 535 | Version Vector 536 | 1983 537 | 538 | 539 | Vector Clock 540 | 1988 541 | 542 | 543 | Bound Version Vector 544 | 2004 545 | 546 | 547 | Interval Tree Clock 548 | 2008 549 | 550 | 551 | Dotted Version Vectors 552 | 2010 553 | 554 |

555 | 556 |

Version Vectors also build on the idea of Lamport Timestamps, but are 557 | specifically meant to track changes to data in distributed systems. They 558 | are also the basis for optimistic replication.

559 |

disadvantages

560 |

Each message sent by a node contains a vector clock that has all the node 561 | names (and their corresponding clock counts) who want to write to the same 562 | field of data.

563 |

This can be a problem since a data structure that can grow to an unbound size 564 | can be a problem in larger networks with more frequent writes. Strategies for 565 | dealing with this are often based on what suits your use-cases best, for 566 | example, two possible solutions are...

567 |

If a network has a finite number of nodes, a message that has reached 569 | all nodes can be considered "complete", could be marked as such and have 570 | it's historic information removed.
571 |
If a network has an acceptable threshold of nodes that once a message has 573 | reached, the message can be considered complete and can then be cleaned up.
574 |

576 | 577 |

578 | 581 |

582 |

583 | 584 |

585 |

implementing dat

586 |

In my spare time I am implementing dat by following this, 587 | this and this as references.

588 |

589 | 590 |

You can follow this post and this github org for updates and information.

591 | 592 |

593 | 596 |

597 |

598 | 599 |

600 |

DTN recap

601 |

History

602 |

Over the last 10 years I've been a part of various meet-ups, IRC channels and p2p 603 | networks that are interested in building distributed systems. In 2015 604 | I pulled together a lot of the leaders and contributors from the projects that I 605 | found interesting for an event we called DTN, aka Data Terra Nemo, 606 | Decentralize The Network, Don't Think Normal. It went well!

607 |

2019

608 |

About 6 months ago Feross, Mikeal and Juan convinced me I 609 | should do another one and it made sense! After all, many of the projects we 610 | discussed in 2015 were just ideas (libp2p), still in the prototyping 611 | phase (dat) or didn't exist at all (filecoin, 612 | patchwork). It happened!

613 |

614 | 615 |

This event works for a few reasons.

616 |

Zero profit. 100% of what we earn gets spent on the conference — 618 | no exceptions. All funds go towards flights, hotels, food, A/V, etc.
619 |
We curate speakers who are hard working, highly motivated, implementers but 621 | also kind, empathetic people. Software is nothing without the people who work 622 | together to make it.
623 |
One of the most important reasons this works is that it's collaborative and 625 | not competitive. We're sharing fundamental ideas. And while we're taking 626 | different approaches to solving some of the same problems, we're all 627 | interested in the best outcome. Things like programming languages are trivial 628 | details.
629 |

631 |
632 | 633 |

What happened?

634 |

Over 2 days about 150 people attended. We recorded about 12 hours of video from 635 | 12 speakers. We had 12 formal talks and several impromptu lightning talks. I'll 636 | be posting the videos online once they are processed.

637 |

We discussed the technical (and some social) challenges in coding 638 | fundamental primitives that help people build a more distributed and 639 | decentralized internet. We shared language agnostic patterns for building 640 | solutions but also many concrete solutions — actual code that illustrates 641 | what problems we're solving today.

642 |

How did you do it?

643 |

I have no idea. It wasn't easy. Unexpected things happened. People are hard to 644 | organize. I'm not an event organizer. I have no idea what I'm doing. I almost 645 | broke-even though. Ultimately it was a lot of fun. And the lesson here is that 646 | anyone can do this. If you want help putting on your own event, I'm happy to 647 | discuss what I know! My handle is heapwolf most places on the internet. If 648 | you're not on any decentralized networks yet, you can even reach me on 649 | twitter.

650 |

What's next?

651 |

Let's do it again! Many of the projects we met to discuss are moving faster, 652 | more people are involved, so let's not make it a 4 year gap this time. Should we 653 | do the same dates? Different Location? Let's talk.

654 |

Thank you!

655 |

If you bought a ticket, you were a sponsor! So I want to thank all the sponsors, 656 | especially those who bought high value tickets. You are helping to build a 657 | better internet for everyone.

658 |

659 | 660 | 661 | 662 | 663 | 664 | 665 | 666 | 667 | 668 |

669 | 670 | 671 |

672 | 675 |

676 |

677 |

678 |

Linux on MacBook Pro

679 |

I recently switched from MacOS to Linux.

680 |

WHY?

681 |

Because learning is fun.

682 |

WHICH LINUX?

683 |

There are thousands of posts about which Linux is the best. There isn't one 684 | right answer. For me it's a distribution that reflects my own values. I like 685 | to keep things simple.

686 |

HELLO, ARCH LINUX.

687 |

688 | 689 |

You can see from the output that Arch Linux is running on MacBook Pro 690 | hardware. That's a screenshot of the terminal running ZSH and TMUX. 691 | MacOS aka Darwin, is a BSD variant, and it's Unix-ish. So transitioning 692 | to Linux is easy and familiar.

693 |
694 | 695 |

GNOME DESKTOP

696 |

There are other desktop options, but Gnome basically gives you an improved 697 | version of MacOS. I added some MacOS icons to make the transition more 698 | seamless.

699 |

700 | 701 |

HARDWARE PREREQUISITES

702 |

Before you do anything, you'll need a USB-C to USB hub. You can buy a cheap 703 | keyboard and mouse to plug into it. You'll also need a USB drive. Get one with 704 | a few gigabytes they are cheap.

705 |

706 | 707 | 708 |

GETTING STARTED

709 |

DISCLAIMER

710 |

This wont be a complete guide. I don't think a complete guide exists. You'll 711 | need to do some searches to find solutions for issues specific to your machine. 712 | Also, this could brick your computer. So don't do any of this unless you really 713 | know how to yolo. If you're a part-time nerd, check out this link.

714 |

OTHER GUIDES

715 |

The two best resources I found were this and this. And of course the 716 | arch wiki was incredibly helpful. I got here by analyzing these resources 717 | and doing some searches.

718 |

BOOT FROM USB

719 |

Start by downloading the ISO from here, pick a link with http! 720 | After you download it, verify the signature.

721 |

You can use this app to make the USB bootable from the downloaded iso.

722 |

Plug everything in and turn your computer on while holding down the command 723 | key. When the menu pops up, select your USB drive and continue booting it. Say 724 | goodbye to MacOS, you're about to install Arch.

725 |

FIRST BOOT

726 |

You'll see some logs and then you'll get dropped into a command line. The first 727 | thing we're going to do is partition and format the disk. Your prompt is going 728 | to look like root@archiso ~ #, the command we're going to run is lsblk.

729 |

root@archiso ~ # lsblk

You'll see a list of bulk devices, mine looks like this but yours will look 730 | different. I don't have a USB drive plugged in. You need to figure out which one 731 | is your storage device and which one is your USB device. You can probably 732 | determine that by the sizes of things.

733 |

NAME        MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
734 | nvme0n1     259:0    0 233.8G  0 disk
735 | |- nvme0n1p1 259:1    0   489M  0 part /boot
736 | |- nvme0n1p2 259:2    0 222.6G  0 part /
737 |  \ nvme0n1p3 259:3    0  10.3G  0 part [SWAP]

PARTITION

738 |

After you figure out your device's name, write it down, we're going to need it 739 | again. Now, let's edit it. For example, if I was to edit my device, I would run 740 | the following command.

741 |

root@archiso ~ # parted /dev/nvme01

You're in the partition editor. Next, destroy everything on the device.

742 |

mklabel gpt

Then with the next two commands create the boot partition.

743 |

mkpart ESP fat32 1049kB 538MB
744 | set 1 boot on

745 |
"Take the size of your hard disk, that you noted yourself earlier, and 746 | subtract the size of your RAM from it. I've got 8GB of RAM, so for SIZE 747 | I've put in: 243GB (251GB minus 8GB)." - Phil
748 |

749 |

You might have to adjust these values, but try something like this to start.

750 |

mkpart primary ext4 538MB 243GB
751 | mkpart primary linux-swap 243GB 100%
752 | quit

FORMAT

753 |

Next we're going to format the partitions. Run lsblk again to see the new 754 | partitions you made, replace foo with your actual device name.

755 |

mkfs.vfat -F32 /dev/foo1
756 | mkfs.ext4 /dev/foo2
757 | mkswap /dev/foo3
758 | swapon /dev/foo3

If you get the error Not enough clusters for a 32 bit FAT!, try increasing the 759 | size of your partition as discussed here.

760 |

BOOTSTRAP

761 |

Next we need to get the OS onto your newly partitioned and formatted device and 762 | configure it.

763 |

MOUNT THE DEVICES

764 |

Let's mount the partitions, replacing foo with your actual device name.

765 |

root@archiso ~ # mount /dev/foo2 /mnt
766 | 
767 | root@archiso ~ # mkdir -p /mnt/boot
768 | root@archiso ~ # mount /dev/foo1 /mnt/boot

INSTALL THE BASE SYSTEM

769 |

Select the right mirror by moving it to the top of the list.

770 |

root@archiso ~ # vi /etc/pacman.d/mirrorlist

Run the pacstrap command to copy the OS onto your device. I needed to install 771 | not only the base and the base-devel packages, but also the wireless 772 | networking packages so that I could get on the Wifi. Later I realized that after 773 | installing the gnome package, network connectivity is handled really well by 774 | the networkmanager package.

775 |

root@archiso ~ # pacstrap /mnt base base-devel dialog wireless_tools netcfg wpa_supplicant

CONFIGURE THE FILE SYSTEM

776 |

Generate your File System Table with this command.

777 |

root@archiso ~ # genfstab -U -p /mnt >> /mnt/etc/fstab

Open it with vi, there are some changes you'll need to make...

778 |

779 |
"Make sure that the line of the ext4 partition ends with a 2, the swap 780 | partition’s line ends with a 0, and the boot partition’s line ends with a 781 | 1. This configures the partition checking on boot." - Phil
782 |

783 |

root@archiso ~ # vi /mnt/etc/fstab

CONFIGURE THE OS

784 |

Now change the root and configure the OS.

785 |

root@archiso ~ # arch-chroot /mnt

PASSWORD

786 |

After you log run arch-chroot, the prompt will change slightly. Type in the 787 | following command to pick a new password for the root user. Write this down.

788 |

[root@archiso /]# passwd

KEYBOARD

789 |

Let's get your keyboard and trackpad working. Use pacman (the Arch package 790 | manager) to install some things.

791 |

pacman -S git kernel-devel dkms

Now edit the keyboard configuration file.

792 |

[root@archiso /]# vi /etc/dracut.conf.d/keyboard.conf

Add the following line.

793 |

add_drivers+="applespi intel_lpss_pci spi_pxa2xx_platform apple-ib-tb"

Now make the system aware of our new modules

794 |

vi /etc/initramfs-tools/modules

Add the following lines

795 |

applespi
796 | apple-ib-tb
797 | intel_lpss_pci
798 | spi_pxa2xx_platform

Now get and build the drivers. If you have a touch-bar, check out the branch 799 | for that using git checkout touchbar-driver-hid-driver after you clone.

800 |

[root@archiso /]# git clone https://github.com/roadrunner2/macbook12-spi-driver.git
801 | [root@archiso /]# cd macbook12-spi-driver
802 | [root@archiso /]# ln -s `pwd` /usr/src/applespi-0.1
803 | [root@archiso /]# dkms install applespi/0.1

There are some tweaks you can do, but at this point you should have a working 804 | keyboard and trackpad!

805 |

reboot

LOCALE

806 |

Open the locale.gen file and uncomment the line with en_US.UTF-8 UTF-8.

807 |

[root@archiso /]# vi /etc/locale.gen
808 | [root@archiso /]# locale-gen

[root@archiso /]# echo LANG=en_US.UTF-8 > /etc/locale.conf
809 | [root@archiso /]# export LANG=en_US.UTF-8

CLOCK & TIMEZONE

810 |

Set the timezone. To get a list of time zones, use timedatectl list-timezones.

811 |

[root@archiso /]# ln -s /usr/share/zoneinfo/Zone/SubZone /etc/localtime

Set the hardware clock.

812 |

[root@archiso /]# hwclock --systohc --utc

KERNEL MODULES

813 |

Add kernel modules by opening or creating the following file.

814 |

[root@archiso /]# vi /etc/modules

Add the following two lines.

815 |

coretemp
816 | applesmc

HOSTNAME

817 |

Set your host name, replace foo with your actual hostname.

818 |

[root@archiso /]# echo foo > /etc/hostname

HOSTFILE

819 |

127.0.0.1   localhost.localdomain foo
820 | ::1         localhost.localdomain foo

NETWORK

821 |

Install and enable the DHCP daemon.

822 |

[root@archiso /]# pacman -S dhcpcd
823 | [root@archiso /]# systemctl enable dhcpcd

BOOTLOADER

824 |

Install EFI tools and use them to install systemd-boot on your boot partition.

825 |

[root@archiso /]# pacman -S dosfstools
826 | [root@archiso /]# bootctl --path=/boot install

[root@archiso /]# vi /boot/loader/entries/arch.conf

Add the following lines and replace foo with the name of your storage device.

827 |

title Arch Linux
828 | linux /vmlinuz-linux
829 | initrd /initramfs-linux.img
830 | options root=/dev/foo2 rw elevator=deadline quiet splash resume=/dev/foo3 nmi_watchdog=0

Now tell the bootloader about your new bootable option.

831 |

[root@archiso /]# echo "default arch" > /boot/loader/loader.conf

Exit arch-chroot, unplug all those USBs and reboot your system.

832 |

[root@archiso /]# exit
833 | [root@archiso /]# reboot

After this you should get dropped into a command line again. But this time you 834 | will be running your new OS and you will have keyboard and mouse support. After 835 | that you may or may not have some more work to do if any of your devices aren't 836 | working (audio either works or can be tricky to get working).

837 |

There are also configuration things to do but it depends on how you want to use 838 | your computer. Most of the code in the official repository is seen by a lot of 839 | eyes. But personally I try to stay away from the AUR if I can. I try to audit 840 | the packages I install.

841 |

You should read the Security section of the Arch Wiki.

842 |

Good luck or congratulations depending on where you are. Hit me up on either 843 | Twitter (@heapwolf) or Freenode IRC heapwowlf if you have questions.

844 | 845 |

846 | 849 |

850 |

851 |

852 |

CODE

853 |

GITHUB

854 |

https://github.com/heapwolf

855 |

PUBLIC KEY

856 |

857 | 858 | ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAACAQC+akORhccxOBxXHKTb3p1CEIUg+/rFxcckARqZ2F3qBOpYzUEPII7aGWwnexgnTLtuaLb0pujFcawne+z/c8mFFnrxc99gG76cRKUsDLcdNXEa3kCBrQG3dXf+J8mIbzIFaabreP0fQAqwI9hbM/SUFwrPbCbfTcZcpn4pgbCSDxLXC0qb1Zh9UkcV4RP3A9r8sy92PrmtUmJTXWdZE043eAXpgerhRkHC38BL2JsmVf6/nOonlHfJiue7nowUaA7bvaA8GGlJRFaNhoFfTYIkMv9pOGaNpTtdK+UpOOO1U8MBlWPI86VOY/0VQj1pZOc8r0hqd3o+ALg9TKtkZKfSgZs4JQRRul4Zvmmc/mJfVJ9x/2rLdSnR19swSkJkyXZh/Q2/4u/Zp1J0+JuWMugLQzkJIfYMEeOm6Z3dk8O3Br1aMA+HPiDLgJjeV5SZm6is+DXdPA9s6CyQ2F3oazD1g1k/34W2p3+dnpAQU6W/73sxzyteUo7FNkGw7uGi7a5Gowmn6VtcoBDyY8QUUduUoCx747DZ61UTFPtesDc2HvPoemH8PYXjduA93CGtdh8ir7HsO9J0uzhC1FqrGskRpjNVZWWo6jERYamnbzYUs4VVy7qxcBy7skAe9JtGq3+MMjt6mvt9doxOAazUiAd/I9YWBCkWk/gwvy1qHa7R/w== hi@hx.ht
859 | 
860 |

861 | 862 |

863 | 866 |

867 |

868 | 869 |

870 |

Curriculum Vitae

871 |

872 | 873 | 874 |

Personal

875 |

Name: Paolo Fragomeni, Software Engineer

876 |

Contact

877 |

Email: paolo@async.ly

878 |

Web: https://hx.ht

879 |

Twitter: https://twitter.com/heapwolf

880 |

Github: https://github.com/heapwolf

881 |

Summary

882 |

I Left MIT in 2010 to co-found Nodejitsu (a PaaS, since integrated with 883 | GoDaddy). Most recently I founded Voltra Co. (entertainment software) which 884 | joined Conductor Lab. In addition to being a technical founder, CTO and engineer 885 | I have worked in the infosec space.

886 |

Expertise

887 |

Computer Science Research. Software Engineering: programming design and 888 | engineering, concurrent and distributed programming, metaprogramming, 889 | functional programming and ECMAScript (Javascript). Key-Value store

890 |

Experience

891 |

CTO, Cofounder at Voltra Co.

892 |

January 2016 - Augest 2018 (2.5 years)

893 |

Voltra Co. was a cloud storage service and set of audio products. Voltra's 894 | desktop and mobile players sync so you can stream your music from anywhere. 895 | The only ad-free fully hi-res platform. On this project I worked with 896 | Electron, JavaScript, Node.js, C++, CSS3, HTML5, Stylus, Jade, Webpack, 897 | React, React-Native, and Amazon Web Services. Voltra joined Conductor Lab 898 | in Augest 2018.

899 |

VP of Engineering at Now Secure

900 |

November 2014 - January 2016 (1 year 3 months)

901 |

Built engineering and security research teams. Coordinated engineering and 902 | research teams. Set technical goals, worked hands on on lots of projects. 903 | On this project I worked primarily with C, C++, JavaScript, Node.js, HTML5.

904 |

Engineer, CTO at Mic

905 |

January 2014 - November 2014 (11 months)

906 |

Hereishow.to joined mic.com where I served as CTO. Built an engineering team 907 | and integrated components of Here Is How into existing products. On this 908 | project I worked with Amazon Web Services, Node.js JavaScript, HTML5, CSS3.

909 |

Engineer, CTO, Cofounder at Here Is How

910 |

November 2012 - January 2014 (1 year 3 months)

911 |

A CMS for technical writing, a web based interface similar to Medium.com. 912 | This project was acqui-hired by mic.com On this project I worked with Docker, 913 | JavaScript, Node.js, Websockets, C, C++, HTML5, CSS3.

914 |

Engineer, CTO, Cofounder at Nodejitsu

915 |

September 2010 - December 2012 (2 years 4 months)

916 |

Co-Founder, Chief Technology Officer. Lots of R&D. Conceptualized and 917 | implemented products that simplify and manage application deployments 918 | for the node.js platform.

919 | 920 |

921 | 924 |

925 |

926 | 927 |

29 |

30 | 31 |

Is web 3.0 here?

32 | 33 | 2021-10-11 34 |

35 | Is bitcoin a ponzi-scheme? Are NFTs a star naming market? Are 36 | ledgers just slow databases? There is some contention among web 37 | developers whether Web 3 is just marketing jargon or an actual 38 | inflection point. 39 |

40 |

41 | This essay examines the legitimacy of the Web 3.0 moniker though the 42 | lens of my 20+ years experience as a computer science researcher, 43 | software engineer, and startup founder. 44 |

45 | 46 |

Before the Web

47 |

48 | Before the Web there were BBSes. A BBS is just a computer running some 49 | software that allows you to connect to it using a phone and a 50 | modem. In the 90's, 51 | there were thousands of these. And a lot of the time a BBS was just a 52 | computer in someone's home. You could find huge lists of BBSes, dial 53 | into them, download files, chat with other users, they even had 54 | federated messages like email. 55 |

56 |

57 | I ran a BBS. I had a few modems, a few phone lines. Only one 58 | connection per phone line was possible, so sometimes all the lines 59 | would be busy, and you'd need to try calling later. At one point I 60 | maxed out on the number of phone lines you're allowed to have in a 61 | residential building. BBSes were the first platforms that focused on 62 | hosting open source code. They were incredible hacker communities. 63 |

64 |

65 | BBSes had clear limitations though. Notably concurrency. Also it 66 | was text based, and users wanted graphical interfaces. There 67 | was no concept of hyper-linking. And toward the end of the 90's, 68 | The Web was emerging. OSes started shipping with a browser 69 | pre-installed, and a shortcut on the desktop by default. It was easy. 70 | You entered an address and saw a website. It was highly concurrent. 71 | It was graphical. It appealed to a every day people as much as hackers. 72 | The turning point was HTTPS, cryptography made it possible to do 73 | business online. BBSes started to die. 74 |

75 |

76 | The thing is... BBS users were extremely skeptical of the 77 | Web — browsers were young. BBS software was mature. 78 | BBSes had advanced, real-time mutli-user interactions! 79 |

80 | 81 |

The Information Super Highway

82 |

83 | No one called the internet "Web 1.0". It had a lot of dumb nicknames. 84 | Very few people were excited about it. And as more people learned 85 | about it, they all described it differently, doubting it in different 86 | ways. 87 |

88 |

89 | I started Web development professionally in the late 90s. Working on a 90 | computer as a job was the absolute opposite of being a rock-star. I 91 | didn't always tell people what I did for a living. Using computers 92 | still came with a social stigma. 93 |

94 |

95 | Everyone was thinking the Web was going to be similar to what we did 96 | on paper, but on a screen — very boring and mostly for business. 97 |

98 |

99 | The early Web was objectively terrible. It was terrible because all 100 | the processing happened on a server. Most people had poor internet 101 | connections, so pages felt huge and slow. If you wanted to do some 102 | kind of interaction, you had to request a completely new page from the 103 | server, even if the smallest amount of data had changed. Also, at that 104 | point, JavaScript was a niche thing that no one wanted to touch. You 105 | had Flash. DOM APIs were slow. Memory leaks were a nightmare. It was 106 | hard to debug and tooling was non-existent. But businesses were optimistic. 107 |

108 |

109 | Toward the end of 2000 people we're going bananas about the Web. The Dot Com Bubble 110 | , about to burst, was rife with get-rich-quick schemes, vaporware, 111 | monumental frauds, and mindless speculaors. 112 |

113 | 114 |

Web 2.0 — A literal breakthough

115 | 116 |

117 | Then in 1999, Microsoft shipped XmlHttpRequest. 118 | And it was an incredible JavaScript API. Because you could fetch 119 | small bits of new data in the background without the whole page 120 | reloading! 121 |

122 |

123 | The kinds of things people started building with xmlHttpRequest were 124 | impressive for the time. For example, Google Maps and Gmail. 125 | 126 | These Web pages were so different and the value they provided was so 127 | dramatically improved, we needed a new way to classify them. The term 128 | we used was "Web 2.0". For years skeptics warned Web 2.0 129 | was just going to be a replay of the Dot Com Bubble. We were surrounded 130 | by new scams — grifters love a catchy moniker. Web 2.0 projects were 131 | just as over valuated as they were in the Dot Com Bubble. Over the next 132 | 5 or 6 years you heard about Web 2.0 everywhere. Recruiters wanted to 133 | see it on your resume. It lost any and all meaning. 134 |

135 |

136 | The take-away is, 1. There was a vital technical capability added, 137 | and 2. equally as important — all browser makers followed suit. 138 |

139 | 140 |

The Web ate a lot of software

141 | 142 |

143 | By 2010 Web development had much better tooling and JavaScript was 144 | evolving. Computing was being distributed out toward the edge and 145 | onto the client, and replacing a lot of traditional desktop software. 146 |

147 | 148 |

149 | The killer value proposition of the Web was how it distributed software. 150 | You entered an address in a browser, and the most up to date software was loaded. 151 | Your intellectual property could be hidden in the cloud, behind a service. 152 |

153 | 154 |

155 | We saw Ruby and Python introduce Web frameworks that made it easy for 156 | developers to ramp up fast and ship things that users wanted. 157 | Github made 158 | sharing code as easy as Twitter made sharing ideas. Node.js came along 159 | and popularized 160 | asynchronous I/O. 161 | I thought Node was so important, I even founded a PaaS for it. 162 |

163 | 164 |

165 | Something else remarkable about the early 10s was The NoSQL vs SQL 166 | wars... this was a pretty famous battle. It made people's blood boil. 167 | After some time though, people realized that this is just about 168 | understanding trade-offs. These days it's easy to say that NoSQL is a 169 | great choice when you have well-known access patterns and most people 170 | won't murder you for saying it. 171 |

172 | 173 |

174 | A lot of truly significant things have pushed the Web forward since 175 | 1999. Some very contentious technologies have come and gone, others 176 | have found their place, and co-exist with what came before them. But 177 | nothing emerged where we all agreed we should declare "Web 3.0". 178 | We were still building client-server Web apps. 179 |

180 | 181 |

The Web started eating itself

182 |

183 | Big companies are the majority contributors of money, time 184 | and resources driving the Web forward. This can be a problem for 185 | lots of reasons. Innovation for example... 186 |

187 |

188 | Google wants the Web to be successful so their ad business can be 189 | successful. But Google isn't incentivized to create the best Web, they're 190 | incentivized to create the Web that is best for their ad-tech. Google 191 | also maintains the incumbent browser with approximately 66% market 192 | share. It's difficult to get market share with a browser. It's also 193 | difficult to contribute to a browser, they're open source, but they're 194 | among the most complex codebases in the world. Understandably, 195 | maintainers want to make it hard to introduce bugs and vulnerabilities. 196 | One does not casually open a Pull Request on Github for a significant 197 | feature. 198 |

199 |

200 | On the other end of the spectrum you have Apple. Apple's Safari has a 201 | relatively large market share at approximately 16%. Yet Apple is 202 | incentivized to make its native platform successful, and they have 203 | strong influence on Webkit's direction with 76.9% of commits 204 | (as of 2021) and the largest number of reviewers. There are a number 205 | of browsers with a sliver of the market share, they appeal to early 206 | adopters, and they are happy to build-in experimental features. 207 |

208 |

209 | Ultimately, the Web, innovation on it, and its potential as a 210 | platform, is limited by the incumbent businesses and banks that 211 | control it. So, where is this all heading? 212 |

213 | 214 |

Decentralization

215 | 216 |

222 | 223 |

224 |

225 |

60s-70s

226 | Mainframes 227 |

228 |

229 |

80s

230 | The PC 231 |

232 |

233 |

90s

234 | Client-Server 235 |

236 |

237 |

90s

238 | P2P 239 |

240 |

241 |

10s

242 | Streaming 243 |

244 |

245 |

20s

246 | Edge 247 |

248 |

249 | 250 |

251 | First we had Mainframes. They filled entire rooms with vacuum tubes 252 | and blinking lights. Entirely centralized for obvious reasons. 253 | Then the Personal Computer 254 | happened, this was entirely decentralized. 255 | Then came Client-Server 256 | — services meant co-dependent consumers, also the cloud is 257 | perfect at hiding IP, bad-code, and even worse behavior — 258 | everything went back toward centralization. 259 | 260 | Meanwhile, p2p networks were becoming 261 | mainstream, Gnutella, FastTrak, and Napster networks had tons of clients, they 262 | were almost as easy to use as browsers. 263 |

264 | 265 |

266 | Streaming services like Spotify and Nextflix were a hard swing 267 | back toward centralization. Luminaries like 268 | the founder of Napster 269 | (who later founded Spotify) realized how hard it was to capitalize on a 270 | market where consumers expected everything for free. Improving the UX 271 | made your average consumer happy. Centralizing and protecting the 272 | content made copyright holders happy. 273 |

274 | 275 |

276 | But here we are today, and we've been talking about the this thing called 277 | the edge 278 | for years. The edge just a term that describes all the compute and 279 | storage that's outside the data-center. 280 | It's growing at an insane pace. The edge has already utterly dwarfed 281 | the compute power of the data-center. Not only that but all these 282 | computers in your pocket, on your desk, in a car, on the airplane, 283 | etc. They are all generating more data, more quickly than ever before. 284 | Streaming, or makeing any kind of requests to and from a server for 285 | computing makes very little sense in terms of cost and performance. 286 |

287 | 288 |

289 | Web 3.0, is ulitmately about (re)decentralization. It's about 290 | addressing the back-pressure of data. And it's pushing us toward 291 | designing hybrid systems that embrace p2p protocols. 292 |

293 | 294 |

295 | Companies are already doing this. A good example is Microsoft. 296 | Think about how many Windows users there are, and how many Windows 297 | updates there are. That's a lot of data. That's a lot of bandwidth. 298 | Cumulativly it's a huge cost. So Windows 10 introduced a feature called 299 | Delivery Optimization, 300 | where updates are downloaded peer-to-peer instead of downloading them 301 | from Microsoft's servers. 302 |

303 | 304 |

Web 3.0 — Peer Pressure

305 |

306 | In the past, p2p protocols were developed by enthusiasts, a tiny group 307 | compared to the number of people working on protocols like http. 308 | Unfortunately this was always a highly fragmented group. But over the 309 | last 10 years services like Github and Gitlab have brought these people 310 | together. 311 | Conference-culture also helped. 312 | There's more people than ever working together in this space, and 313 | now we're starting to see a large number 314 | of high-quality solutions for problems like 315 | Peer Discovery 316 | and Hole Punching. 317 | We're seeing developer friendly frameworks like libp2p, 318 | hypercore-protocol, 319 | and hyperswarm. 320 | We're even seeing a nice decentralized competitor to Github called 321 | Radicle. 322 |

323 | 324 |

325 | But Web 3.0 isn't just about p2p projects becoming viable, or easier 326 | for the average developer to build with. There's potential for building 327 | things differently, discovering new value, and escaping current flaws. 328 | Naval Ravikant summerized his 329 | thoughts rather nicely, ironically on Twitter. 330 |

331 | 332 |

333 | Web 2.0: Users are the data, corporations are the platform. The code is closed. 334 | Web 3.0: Users own the data, contributors own the platform. The code is open. 335 |

336 | 337 |

338 | There are going to be things branded as Web 3.0 that we laugh at. 339 | There are going to be a lot of platforms that claim their platform is 340 | the base primitive for all things to be built on (these are the most 341 | suspicious, I won't name names). There's going to be a ton of 342 | vaporware. Loads of scams and get-rich-quick schemes. But this is 343 | how it's always been. 344 |

345 |

Fortunately, at it's core, Web 3.0 is a push to decentralize. To 346 | incrementally embrace p2p architecture, to redistribute ownership and 347 | execution. Web 3.0 offers us a chance to improve availabilty and 348 | reduce costs. And with this, there's a lot to be cautiously optimistic 349 | about. 350 |

351 |

352 |

353 | 354 |

C++20 Modules

355 | 356 | 2021-04-29 357 |

358 | I started digging into C++ 20 Modules. 359 | If you're not familar, 360 | here 361 | is a good introduction. 362 | Compared to other languages like Rust, I found official references and examples sparse or incomplete. 363 | Microsoft has a short intro. I found these 364 | Vector of Bool 365 | posts insightful, although the writer has seemed to receive C++ Modules with a high degree of criticism. 366 | I find the Clang docs unclear about a 367 | number of important things. Things are slightly different with GCC. 368 |

369 | First of all, with Clang 12, you need to opt-in to modules with -fmodules. I'm not sure why 370 | -Xclang -emit-module-interface is a front-end option still. It seems obscure, 371 | you can only find it from running clang -cc1 --help. But you need it because you're expected to 372 | precompile the module (I made a gist that demonstrates the work flow). Then when you build the 373 | file that imports the module, you need to specify where the prebuilt module is located. 374 |

375 | It's awkward that the compiler wants to know the module names. Your compiled module's file name 376 | must match the module name or you can provide a special flag that maps the module name to an arbitrary file 377 | (ie -fmodule-file=name=./file.pcm). 378 | 379 | 380 | 381 |

382 | C++ Modules are literally an after-thought, and they have the DX of one. They get 383 | a lot more complex than this, I will spare you the rather painful details. Honestly, 384 | I'm not sure if they're going to appeal to anyone who's got a large C++ codebase. 385 | But for new projects they seem to be an improvement. 386 |

387 | 388 | 389 | 390 | 391 |

392 | 393 |

Falsifiability

394 | 395 | 2020-11-06 396 |

402 | Verifying the claim "All swans are white" would require assessment of all swans, 403 | which is not possible, the single observation of a black swan is sufficient to falsify it. 404 | 405 | Citation — Falsifiability 406 | 407 |

408 |

409 | 410 |

411 | 412 |

Small-world networks

413 | 414 | 2020-05-31 415 |

416 | The Watts-Strogatz model of the small-world network 417 | have a large clustering coefficient 418 | and paths from node a to b that are on average short. 419 |

420 | This means the typical number of steps between two randomly chosen nodes grows 421 | proportionally to the logarithm of the number of nodes in the network. The simplicity 422 | of this approach makes it a nice starting point for decentralized search. 423 |

424 |

425 | "It is possible to prove that in the model of a d-dimensional lattice with uniformly 426 | random shortcuts, no decentralized algorithm can find short paths [...]. Exploring 427 | further, though, we find that a subtle variant of the Watts-Strogatz network 428 | will in fact support efficient search: Rather than adding the long-range shortcuts 429 | uniformly at random, we add links between nodes of this network with a probability that 430 | decays like the dth power of their distance (in d dimensions)." 431 | 432 | Jon Kleinberg — The Small-World Phenomenon and Decentralized Search 433 | 434 |

435 |

436 | 480 | 481 | 482 | 483 |

484 | 485 |

486 | 487 |

A blockchain is just a data structure

488 | 489 | 2020-05-30 490 |

491 | It has no moving parts. It's just a list of items. And each item has 492 | some properties. The defining characteristics though, are that each 493 | item has 1. a property that acts like a finger print, and 2. a property 494 | that points to the item that came before it. A secure 495 | hash 496 | function creates the finger print using the item's properties as input. 497 | A correct and complete blockchain can be implemented in fewer than 50 lines 498 | of code. 499 |

500 | A blockchain is most likely the least interesting part of any system that 501 | is built with one. Networking, consensus and replication, 502 | for example, are discrete and non-trivial problem domains that can form a 503 | foundation for maing a blockchain useful. 504 |

505 | 506 |

507 | 508 |

Cut And Paste Cults

509 | 510 | 2020-05-27 511 |

512 | The most remarkable feature of most frameworks is their ability to manage mediocre developers. 513 |

514 | 515 |

516 | 517 |

On Minimalism

518 | 519 | 2020-05-01 520 |

521 | Constraints provide meaning. 522 |

523 | 524 |

525 | 526 |

Clocks

527 | 528 | 2020-04-02 529 |

530 | Some papers on logical and vector clocks. 531 |

532 | 533 | Lamport Timestamp 534 | 1978 535 | 536 | 537 | Version Vector 538 | 1983 539 | 540 | 541 | Vector Clock 542 | 1988 543 | 544 | 545 | Bound Version Vector 546 | 2004 547 | 548 | 549 | Interval Tree Clock 550 | 2008 551 | 552 | 553 | Dotted Version Vectors 554 | 2010 555 | 556 |

557 |

558 | 559 |

560 | 561 |

Lamport Timestamps

562 | 563 | 2020-03-01 564 |

565 | Interactive logical timestamps. 566 |

567 |

568 |

569 |

570 | 571 |

572 | Fire node A 573 | Fire node B 574 |

575 |

576 | 577 |

${dir}

Personal

Summary

Expertise

Experience

Founder, Engineer, CEO — Socket Supply Co.

Technologist — Passport Capital

CTO, Cofounder — Voltra Co.

Organizer — Data Tera Nemo

VP of Engineering — Now Secure

Engineer, CTO, Cofounder — Here Is How

Engineer, CTO, Cofounder — Nodejitsu

a little class for combining behavior, layout and style

2018-7-30

2019-7-3 Update

What are javascript then-ables

Await-able Classes

Await-able Objects

Await-able object factories

Async then-ables

preamble

synopsis

description

illustrated lamport timestamp

problem

solution

Clocks vs. Logical Clocks

Example

vector clocks

problem

solutions

sending messages

receiving messages

handling concurrent messages

research timeline

disadvantages

implementing dat

DTN recap

History

2019

What happened?

How did you do it?

What's next?

Thank you!

Linux on MacBook Pro

WHY?

WHICH LINUX?

HELLO, ARCH LINUX.

GNOME DESKTOP

HARDWARE PREREQUISITES

GETTING STARTED

DISCLAIMER

OTHER GUIDES

BOOT FROM USB

FIRST BOOT

PARTITION

FORMAT

BOOTSTRAP

MOUNT THE DEVICES

INSTALL THE BASE SYSTEM

CONFIGURE THE FILE SYSTEM

CONFIGURE THE OS

PASSWORD

KEYBOARD

LOCALE

CLOCK & TIMEZONE

KERNEL MODULES

HOSTNAME

HOSTFILE

NETWORK

BOOTLOADER

CODE

GITHUB

PUBLIC KEY

Curriculum Vitae

Personal

Contact

Summary

Expertise

Experience