Rolandt

Doug Johnson, The Blue Skunk Blog, Jun 28, 2019

Recently a U.S. presidential candidate announced a proposal to pay off all student debts. What surprised me was not the announcement so much but the wave of teachers and professors denouncing the idea. There is this myth, you see - and it is a myth - that people won't value something if it's free. Really? The next time you breathe, tell yourself that. The next time you make a friend, tell yourself that. The next time you fall in love, tell yourself that. The next time you wank down an open road, tell yourself that. Tell yourself that you value none of these things because they're free. See, here's the thing about this sort of argument - they're always about other people. Of course, I would have valued my education no matter what, but other people, they're of lesser moral character (insert apocryphal example here) and they would abuse it. Bah.

Web: [Direct Link] [This Post]

Prion, just a thought here. With my Chromebook I use Google docs offline all the time. It is not necessarily dependent on being connected to the net. One good thing about getting my Chromebook in January is that I took a serious look at Google Docs, and so much prefer it to Word. A key advantage for me is being able to have more than one window open at a time.

Daly

Excellent piece by Dieter Bohn at The Verge following news of Jony Ive’s coming departure from Apple:

While Apple might have a good story about having been founded in a garage, the true founding myth of Apple is the myth of genius. You know the fable, which has the benefit of also being true. When Steve Jobs was in charge, Apple made amazing things: the Apple computer, the Mac. Jobs not in charge: the very bad ‘90s with Scully and the Newton. Jobs back in charge: the renaissance, the iPod, the iPhone.

After Steve Jobs, that mantle was passed to Jony Ive. And he quietly (quite literally) took it. It was important to our concept of Apple that there be a single, discerning decision maker. Somebody uncompromising about quality. Somebody with very good taste. A capital G Genius.

Bohn makes the case, based on solid evidence from other sources, that Apple has operated for years without being driven by a singular “genius” but rather a collaborative, highly-capable team – and while that seems to have been more true than ever lately, to a degree it’s always been the case. In spite of the mythos surrounding Steve Jobs, responsibility for Apple’s best work falls not just on his shoulders, but on that of the team he was surrounded by.

Ive’s absence will certainly be felt, but the hole he leaves is likely much smaller than his “legend” would imply. As Bohn remarks, “we should stop thinking of Apple as the singular expression of one person’s genius. History has moved beyond the Great Man theory, and so too should our ideas about how Apple operates.”

→ Source: theverge.com

Time flies and I am always astounded that I’ve been writing this blog since 2005. Over the years, quite a number of posts have accumulated. As my current WordPress profile bundles them in pages of 10 posts, I’m now on page 255 and thus close to an 8-bit overflow. If you don’t get it, you are probably too young and never lived through the 8-bit home computer era ^^

Prion wrote:

>As a researcher using a Mac I have always asked myself why we keep
>sending around Word documents.

Any writing app can be used as long as everyone on the team has it, is conversant enough with it to write the required document in a base format for final submission, and follows a common approach to using it for the collaborative writing. Word checks the first two boxes most often. You and the team just have to assure that the third guideline is met. In this regard, I have found the best approach is to ask everyone up front to respect a few guidelines:

1) During the writing, use cite keys in the document (carry over from LaTeX) rather than a numbering approach. For example, cite{Barns2019, Jones2017, Smith2006a} references documents that are tagged by the representative keys.

2) Assign one person on the team to manage the citations. He or she handles them during the writing and compiles them into the final document at the end. Everyone sends their citations to the one person in a library format such as BibTeX or RIS and cross-referenced using the cite key standards.

3) Assign one person as the final arbitrator of any conflicts in proposals for editing.

4) Use a common cloud storage for the document; avoid email. This keeps documents from getting misplaced among other things.

5) Be religious about versioning the files at every step and about the notations used for to denote different file versions. For example, a file titled document v03 jjw,cn denotes (release) version 3 of the document that has recently been annotated by jjw followed by cn.

6) Establish up front that you will set defined points in time (e.g. on a weekly basis every Friday) to compile the document to a "release version". This avoids that annotations accumulate beyond reasonable control.

>Personally, I suspect that a third reason that no one speaks about is
>the real death blow for online collaborative writing in academia: They
>make it bloody obvious who contributed how much and when. There is no
>hiding behind "Oh, I did write a whole paragraph but forgot t send it
>around" followed by a hectic writing session in the afternoon.

Hence the reason for agreeing on guideline 6. OK, so you wrote that paragraph this morning but did not post it before our review session. What you wrote will have to be added to the next release.

Otherwise, the debates on who added what and when happened even in the days of the old fashion method of sending hard copies around. So, no this is not a death blow to online collaborative writing, it is a death blow to doing good team work during the writing process itself.

>People
>feel uncomfortable when their writing activity is visible to others in
>real time, if only in theory.

This is perhaps closer to the truth as far as why on-line, interactive collaborative writing is not gonna happen for bigger documents. In such a case, I don't want or indeed need someone looking over my shoulder while I write. I prefer to hack about in the document during my own time away from the internet, post my edits for review, and then have a video conference or phone call to discuss the proposed inputs and reach a consensus on the release version.

The opposite is perhaps true of memo style documents, where immediate and dynamic feedback is demanded during the editing in order to create a document that has a true consensus of all involved.

>As you may guess, Word and Bookends (yay, I I did not give in
>completely) it is for me still. If anyone has found a solution that I
>have overlooked please let me know.

I prefer LaTeX for my own work but accept that others prefer Word. I won't ever do the work to compile citations or to generate in-line figures in Word documents. I have at times taken an entire Word document over to LaTeX format to create a final submission.

I am transitioning to Bookends for my own personal citation management. I would have continued to used Papers but ... it is just taking too long in its transition to Papers/ReadCube (among other reasons) to remain a viable team player. When I can control the choice, I demand the use of a cloud storage option for citation management with teams. By example, I require each of my students to create their own private group in Mendeley and share it with me. For each report or publication that they write, I require them to create a Mendeley folder in their shared drive and stores all of the associated citations to that report or publication.

More on Evernote's struggles in today's NY Times:

A Unicorn Lost in the Valley, Evernote Blows Up the ‘Fail Fast’ Gospel

https://www.nytimes.com/2019/06/28/business/evernote-what-happened.html

Daly de Gagne wrote:
Dr Andus, if you get a chance, and perhaps in a new thread, I would be
>quite interested in how you use Dynalist for research as a Zettelkasten.
>Now that I am primarily using a Chromebook, and loving it, I have
>wondered about using Dynalist more intentionally for notes and bibdata.
>I would use date and time as a unique identifier - 2019 06 23 2045h,
>and appropriate tags.

My system is quite simple. I have a separate file for Zettelkasten, which only holds my reading notes.

Each entry's title starts with a date identifier: 2019-06-29 15:25 Author (pub. year) descriptive title of entry

Then in the note section of the entry there is either a quote or my summary of the point.

I may have an additional "NOTE" section for some meta observation.

Then comes a "See also" section where I add the links to other already existing Zettels in the system.

Then comes the bibliographic info, and URL link to the source, if a web source. Plus "Date accessed:"

And at the bottom I list all the categories in a single row: @Einstein if it relates to a human, and #topic for all other categories.

I also use a few CSS style sheets with the Stylus extension in Chrome, to enhance the visual presentation. Row separator creates the "Zettel" (or card) effect, and the Darknest Hour colours in URLs and tags, so they can be more easily spotted. Coloured Pills is for marking out Zettels for further attention.

Darknest Hour v.1.2
https://userstyles.org/styles/141020/darknest-hour-v-1-2

Dynalist Row Separator
https://userstyles.org/styles/134412/dynalist-row-separator

Dynalist coloured pills
https://userstyles.org/styles/161369/dynalist-coloured-pills

The Create Link Chrome extension is also helpful for pasting in Markdown-formatted links of web sources.

Jeffery Smith wrote:
>What are the alternatives that have
>Evernote's functionality?

Has anybody tried to replace Evernote by Notion? I've just noticed their promo: 'Tired of Evernote? Say hi to Notion'
https://www.notion.so/evernote
and wonder whether it is something worthy of trying?

“The original is unfaithful to the translation.” - Jorge Luis Borges

This makes me want to dance.

Making some changes, because Medium. My work with the On The Horizon project – researching...

I started supporting the IndieWeb with a small monthly contribution. You can do too if you want.

During the introductions at IndieWeb Summit, Malcolm Blaney showed his site Unicyclic, which is set up as an IndieWeb reader. He has an overview page of recent IndieWeb related postings. To have your own postings incorporated, you need to follow his IndieWeb page and send it a WebMention. Upon receiving the WebMention it will look for a feed to follow on your own site. That sounded like a fun thing to try.

Read Podcast number 300: “Adversarial Interoperability: Reviving an Elegant Weapon From a More Civilized Age to Slay Today’s Monopolies” | Cory Doctorow's craphound.com by Cory Doctorow | Cory Doctorow's craphound.com

Adversarial interoperability is the consumer’s bargaining chip in these coercive “negotiations.” More than a quarter of Internet users have installed ad-blockers, making it the biggest consumer revolt in human history. These users are making counteroffers: the platforms say, “We want all of your data in exchange for this service,” and their users say, “How about none?” Now we have a negotiation!

Adversarial Interoperability, a useful concept to keep in mind. In part the IndieWeb is a form of this, as it offers a way of staying outside walled gardens, while still being able to pass messages back and forth through its gates (i.e. API’s), through POSSE / sometimes PESOS. Though some platforms, Facebook actually, made the ‘counter offer’ of switching off their API’s. Twitter similarly has been on a path of absorbing into itself all kinds of apps (e.g. Tweetdeck) that were independent parts of the ecosystem growing on Twitter’s API, and increasing the threshold for access to the API.

Interoperability is a core value to maintain. Use it.

For whatever it's worth, Alex Jenter has contributed 47 posts to the CintaNotes discussion forum since the beginning of April -- the most recent one dated yesterday. I'm not at all persuaded that he has abandoned CN; my one complaint, as I said earlier, is that the development proceeds at a snail-like pace. But I learned long ago that you sometimes have to put up with that when you are dealing with one- or two-person operations like CN.

I forgot to mention, incidentally, that CN does not offer an iOS version, but it will sync with Simplenote on my iPhone if I wish or -- my preferred solution -- I can very easily export an HTML file and then import that (from Dropbox) into a synced folder in the Documents app my phone. This file is of course static but is attractively formatted and easily searchable. Bringing it up to date from time to time takes only a few keystrokes.

Because I was raised in an era when we were willing to conceive of mobile devices as servers as much as clients, it’s always bothered me that the ability to mount an Android phone’s storage on a desktop PC or Mac disappeared when Google removed USB Mass Storage Mode from its operating system.

My phone is a powerful sensor array, it can have a Linux command line, making it a programmable mobile powerhouse. Not being able to easily transfer files to and from its storage is disabling.

But there’s a solution for this.

On the Phone

For greatest ease, assuming your version of Android supports it, assign your phone a static IP address on your wifi network; on Android Pie you’ll find this under the “Advanced” settings for your wifi connect:

Install Termux, either from Google Play or from F-Droid.

Once it’s installed, start the app, and start an SSH server with:

sshd

On the Mac

For greatest ease, add an entry in /etc/hosts for your phone like this:

192.168.2.8	phone

You can do this with:

sudo nano /etc/hosts

Next, assuming you already have Homebrew installed (look here if you need to install it):

brew cask install osxfuse
brew install sshfs

This will install SSHFS, which you’ll use on your Mac to mount the Android storage.

Once it’s installed, create a directory where you’d like to mount the phone storage, and do the mount:

mkdir motog7
sshfs phone:/storage/emulated/0 ~/motog7 -o volname=motog7 -p 8022

In that sshfs command:

• phone is the name I assigned to my Android phone’s static IP address in /etc/hosts
• /storage/emulated/0 is the Android path I want to mount on my phone
• ~/motog7 is the directory I created on my Mac for the mount
• -p 8022 sets the SSH port to use for the mount as 8022, which is what Termux uses by default

You might get a warning dialog “System Extension Blocked” when you attempt the mount; you can allow this to proceed under System Preferences > Security & Privacy.

Use your new superpowers…

Assuming all went according to plan, you now have the Android’s storage mounted on your Mac.

It will show up in the Finder:

And it will be accessible from the command line:

macmini:~ peter$ ls ~/motog7
Alarms		Download	Recordings	bluetooth
Android		Movies		Ringtones	osmand
BROTHER		Music		Signal		osmdroid
Cardboard	Notifications	Telegram	osmtracker
DCIM		Pictures	Vespucci
Documents	Podcasts	alt_autocycle

You can treat it like another disk drive.

To Unmount

To umount the phone:

umount ~/motog7

You may find that the mount kills itself if the Termux app on your phone is terminated by Android.

GitHub, Google, Jun 28, 2019

Something interesting from Google. "The goal of Project Oak is to create a specification and a reference implementation for the secure transfer, storage and processing of data." The thinking here is that while we can lock down and encrypt data when we're transferring it, when it gets into the recipient computer it is unlocked and any system on the computer can access it, making it insecure. So this project posits what are called "enclaves" in recipient computers where all the data processing will happen in an area locked away from the rest of the computer. "Enclaves protect data and code even from the operating system kernel and privileged software, and are intended to protect from most hardware attacks." Of course, what enclaves also protect data from are computer users. They are the modern incarnation of 'trusted computing' - that is, a part of your computer that you can't access, and hence is 'trusted' by content providers (like the purveyors of digital rights management solutions). Beware the Oak bearing gifts.

Web: [Direct Link] [This Post]

There have been hints (and there will be no end of “no more TouchBar/better keyboard” jokes) but this is both interesting timing and probably long overdue.

Like Gruber points out, Apple’s product pipeline and planning cycle is long enough for the effects to be latent, but it is still something to worry about in this age of relentless business operations Tim has brought to the company.

After all, if both your software and hardware innovation starts petering out, you lose your competitive edge.

Or, like Steve would have probably said, your soul.

Prion wrote:
>As a researcher using a Mac I have always asked myself why we keep
>sending around Word documents.

Thanks for this excellent wrap-up of options. I am on Windows and Linux and the options I see are just about the same.

While Word may be built for Windows, it is definitely _not_ built for collaborative editing. In theory one can avoid multiple conflicting versions (usually a result of sending the 'latest' version around via email!) by integrating Word with Dropbox, but this only means getting a warning when someone else is working on the same file as you are, not integrating both (or more) concurrent users' changes.

At our office we have since several years relied on the Google suite for collaborative editing. I have also managed to convince external collaborators to work on single Google documents (or spreadsheets) when acting as the coordinator --thankfully, responsibility usually implies choice of method and tools. When someone else is leading, it is usually a shared Dropbox folder at best (at worst, SharePoint!)

We have used Atlassian Confluence extensively for specific projects (where the structuring of many documents is required). While not ideal for concurrent editing of the same document ('page' in Confluence speak) either, it can work well by assigning separate parts of the overall structure to the relevant users. It is also very extendable and can produce a large variety of outputs starting from the same content. I've mentioned this as a relevant reference in the past: https://contentmarketinginstitute.com/2015/02/technology-behind-language-of-content-strategy/
While some programming may be required for extensive customisation, actual users only see a quite friendly WYSIWYG interface.

Some years ago, there was an interesting offering specifically for collaborative writing, Quip: https://quip.com/about/product This has now been acquired by Salesforce and taken a more specific direction (not sure if the direction was the result of the acquisition, or the reason for it).


>Personally, I suspect that a third reason that no one speaks about is
>the real death blow for online collaborative writing in academia: They
>make it bloody obvious who contributed how much and when. There is no
>hiding behind "Oh, I did write a whole paragraph but forgot t send it
>around" followed by a hectic writing session in the afternoon. People
>feel uncomfortable when their writing activity is visible to others in
>real time, if only in theory.

Oh, this happens in academia as well? I thought it was mostly an issue in proposal and report writing ;)

Matthias Ott, Jun 28, 2019

I think I might do this sometime in the future (it's 4:31 on the day before my summer vacation so I'm not starting anything just right now). " Many personal websites, including this one here, have an “about” page. It’s a page that tells you something about the background of a person or about the services provided. But what this page often doesn’t answer... is what this person really is up to at the moment. A page that answers questions like: What are you focused on at this point in your life? What have you just started working on that excites you like nothing else? Did you just move to a new town? Did you start a new career as a Jengascript wrangler? To answer all those questions, Derek suggests to create a 'now page'." It's a neat idea. Here are some examples of sites with Now pages. Via Chris Coyier.

Web: [Direct Link] [This Post]

While Toronto and Vancouver didn’t make the top of Google Maps’ most crowded transit lines in the world, the tech giant has listed which stops are the busiest in each of the Canadian cities.

In Toronto, Google Maps indicated that routes 68, 501 and 103 are the busiest transit routes.

On route 68, the most crowded stops are, Warden Ave. at Manhattan Dr., Warden Ave at Marble Arch Cres., and Murray Glen Rd.

On route 501, the most crowded stops are Queen St. East at Sudbury St., Queen St. West at Bathurst St., and Queen St. East at Brooklyn Ave.

On route 103, the most crowded stops are Hurontario St. at Ray Lawson Blvd., Brampton Gateway Terminal, and Sir Lou.

In Vancouver, Google Maps said routes 319, 14 and 49 are the most crowdest routes.

On route 319, the most crowded stops are Northbound Scott Rd. at 75A Ave. Westbound 72 Ave. at 126 St., and Westbound 72 Ave. at 122A St.

On route 14, the most crowded stops are Northbound Granville St. at W Broadway, Eastbound W Broadway St. at Balaclava St., and Eastbound W Broadway at Fir St.

On route 49, the most crowded stops are Westbound E 49 Ave at Fraser St., Westbound E 49 Ave at Elgin St., and Westbound E 49 Ave at Wales St.

According to Google Maps, the top ten most crowded transit lines (including subways, metros, commuter rails and other heavy transit) include:
1. Buenos Aires Urquiza Line
2. Sao Paulo Line 11 (Coral)
3. Buenos Aires Line A
4. Sao Paulo Linha 8 Diamante
5. Paris Line 13
6. Buenos Aires Line C
7. Tokyo Chuo Line
8. Sao Paulo Line 9 (Emerald)
9.Tokyo Nippori-Toneri Liner
10. New York L train

According to the blog post, the “crowdedness predictions come from optional feedback directly from the people who use Google Maps.”

Individuals receive notifications asking them about how crowded their subway train or bus ride was after navigating in transit mode, Google Maps said.

Results were taken from users from October 2018 to June 2019 during peak commuting hours (6am – 10 am) and “identified which lines had the highest number of crowdedness reports.”

Source: Google Maps

The post Google Maps reveals most crowded transit lines around the world appeared first on MobileSyrup.

The post The five best times Jony Ive was a disembodied voice in an Apple video appeared first on MobileSyrup.

This is wrong:

Because I’m not blocking ads. I’m blocking tracking.

In fact I welcome ads—especially ones that sponsor The Washington Post and other fine publishers. I’ll also be glad to subscribe to the Post once it stops trying to track me off their site. Same goes for The New York Times, The Wall Street Journal and other papers I value and to which I no longer subscribe.

Right now Privacy Badger protects me from 20 and 35 potential trackers at those papers’ sites, in addition to the 19 it finds at the Post. Most of those trackers are for stalking readers like marked animals, so their eyeballs can be shot by “relevant,” “interest-based” and “interactive” ads they would never request if they had much choice about it—and in fact have already voted against with ad blocking, which by 2015 was already the biggest boycott in world history. As I point out in that link (and Don Marti did earlier in DCN), there was in that time frame a high correlation between interest in blocking ads and interest (surely by the ad industry) in retargeting, which is the most obvious evidence to people that they are being tracked. See here:

Tracking-based ads, generally called adtech, do not sponsor publications. They use publications as holding pens in which human cattle can be injected with uninvited and unwelcome tracking files (generally called cookies) so their tracked eyeballs can be shot, wherever they might show up, with ads aimed by whatever surveillance data has been gleaned from those eyeballs’ travels about the Net.

Real advertising—the kind that makes brands and sponsors publications—doesn’t track people. Instead it is addressed to whole populations. In doing so it sponsors the media it uses, and testifies to those media’s native worth. Tracking-based ads can’t and don’t do that.

That tracking-based ads pay, and are normative in the extreme, does not make right the Post‘s participation in the practice. Nor does it make correct the bad thinking (and reporting!) behind notices such as the one above.

Let’s also be clear about two myths spread by the “interactive” (aka “relevant” and “interest-based”) advertising business:

1. That the best online advertising is also the most targeted—and “behavioral” as well, meaning informed by knowledge about an individual, typically gathered by tracking. This is not the kind of advertising that made Madison Avenue, that created nearly every brand you can name, and that has sponsored publishers and other media for the duration. Instead it is direct marketing, aka direct response marketing. Both of those labels are euphemistic re-brandings that the direct mail business gave itself after the world started calling it junk mail. Sure, much (or most) of the paid messages we see online are called advertising, and look like advertising; but as long as they want to get personal, they’re direct marketing.
2. That tracking-based advertising (direct marketing by another name) is the business model of the “free” Internet. In fact the Internet at its base is as free as gravity and sunlight, and floats all business boats, whether based on advertising or not.

Getting the world to mistake direct marketing for real advertising is one of the great magic tricks of all time: a world record for misdirection in business. To help explain the difference, I wrote Separating Advertising’s Wheat From Chaff, the most quoted line from which is “Madison Avenue fell asleep, direct response marketing ate its brain, and it woke up as an alien replica of itself.” Alas, the same is true for the business offices of the Post and every other publisher that depends on tracking. They ceased selling their pages as spaces for sponsors and turned those spaces over to data vampires living off the blood of readers’ personal data.

There is a side for those publishers to take on this thing, and it’s not with the tracking-based advertising business. It is with their own moral backbone, and with the readers who still keep faith in it.

If any reporter (e.g.@CraigTimberg , @izzadwoskin, @nakashimae ‏and @TonyRomm) wants to talk to me about this, write me at doc at searls.com or DM me here on Twitter.* Thanks.

Bonus link (and metaphor)

*So far, silence. But hey: I know I’m asking journalists to grab a third rail here. And it’s one that needs to be grabbed. There might even be a Pulitzer for whoever grabs it. Because the story is that big, and it’s not being told, at least not by any of the big pubs. The New York Times‘ Privacy Project has lots of great stuff, but none that grabs the third rail. The closest the Times has come is You’re not alone when you’re on Google, by Jennifer Senior (@JenSeniorNY). In it she says “your newspaper” (alas, not this one) is among the culprits. But it’s a step. We need more of those. (How about it, @cwarzel?)†

[Later…] We actually have a great model for how the third rail might be grabbed, because The Wall Street Journal wrestled it mightily with the What They Know series, which ran from 2010 to 2012. For most of the years after that, the whole series, which was led by Julia Angwin and based on lots of great research, was available on the Web for everybody at http://wsj.com/wtk. But that’s a 404 now. If you want to see a directory of the earliest pieces, I list them in a July 2010 blog post titled The Data Bubble. That post begins,

The tide turned today. Mark it: 31 July 2010.

That’s when The Wall Street Journal published The Web’s Gold Mine: Your Secrets, subtitled A Journal investigation finds that one of the fastest-growing businesses on the Internet is the business of spying on consumers. First in a series. It has ten links to other sections of today’s report.

Alas, the tide did not turn. It kept coming in and getting deeper. And now we’re drowning under it.

† I did hear from Charlie Warzel (@cwarzel), who runs the Privacy Project series at the Times , and assured me that they would be covering the issue. And (Yay!) it did, with I Visited 47 Sites. Hundreds of Trackers Followed Me, by Farhad Manjoo (@fmanjoo). This was followed by critique of that piece titled Privacy Fundamentalism, by Ben Thompson in Stratechery. I responded to both with On Privacy Fundamentalism. So check those out too.

The r-project.org domain had some temporary technical difficulties this week (2019-29) that made reaching R-related resources problematic for a bunch of folks for a period of time. Incidents like this underscore the need for regional and network diversity when it comes to ensuring the availability of DNS services. That is, it does no good if you have two DNS servers if they’re both connected to the same power source and/or network connection since if power goes out or the network gets wonky no client will be able to translate r-project.org to an IP address that it can then connect to.

I’m not at-keyboard much this week so only had time to take an external poke at the (new) r-project.org DNS configuration late yesterday and today before the sleepyhead vacationers emerged from slumber. To my surprise, the r-project.org current DNS setup allows full zone transfers, which means you can get the full “database” of r-project.org DNS records if you know the right incantations.

So, I wrote a small R function wrapper for the dig command using {processx}. Folks on the legacy Windows operating system are on your own for getting a copy of dig installed but users of proper, modern operating systems like Linux or macOS should have it installed by-default (or will be an easy package manager grab away).

Wrapping dig

The R-wrapper for the dig command is pretty straightforward:

library(stringi) # string processing
library(processx) # system processes orchestration
library(tidyverse) # good data wrangling idioms

dig <- function(..., cat = TRUE) {

  processx::run(
    command = unname(Sys.which("dig")), 
    args = unlist(list(...)),
  ) -> out

  if (cat) message(out$stdout)

  invisible(out)

}

We expand the ellipses into command arguments, run the command, return the output and optionally display the output via message().

Let’s see if it works by getting the dig help:

dig("-h")
## Usage:  dig [@global-server] [domain] [q-type] [q-class] {q-opt}
##             {global-d-opt} host [@local-server] {local-d-opt}
##             [ host [@local-server] {local-d-opt} [...]]
## Where:  domain    is in the Domain Name System
##         q-class  is one of (in,hs,ch,...) [default: in]
##         q-type   is one of (a,any,mx,ns,soa,hinfo,axfr,txt,...) [default:a]
##                  (Use ixfr=version for type ixfr)
##         q-opt    is one of:
##                  -4                  (use IPv4 query transport only)
##                  -6                  (use IPv6 query transport only)
##                  -b address[#port]   (bind to source address/port)
##                  -c class            (specify query class)
##                  -f filename         (batch mode)
##                  -i                  (use IP6.INT for IPv6 reverse lookups)
##                  -k keyfile          (specify tsig key file)
##                  -m                  (enable memory usage debugging)
##                  -p port             (specify port number)
##                  -q name             (specify query name)
##                  -t type             (specify query type)
##                  -u                  (display times in usec instead of msec)
##                  -x dot-notation     (shortcut for reverse lookups)
##                  -y [hmac:]name:key  (specify named base64 tsig key)
##         d-opt    is of the form +keyword[=value], where keyword is:
##                  +[no]aaonly         (Set AA flag in query (+[no]aaflag))
##                  +[no]additional     (Control display of additional section)
##                  +[no]adflag         (Set AD flag in query (default on))
##                  +[no]all            (Set or clear all display flags)
##                  +[no]answer         (Control display of answer section)
##                  +[no]authority      (Control display of authority section)
##                  +[no]besteffort     (Try to parse even illegal messages)
##                  +bufsize=###        (Set EDNS0 Max UDP packet size)
##                  +[no]cdflag         (Set checking disabled flag in query)
##                  +[no]cl             (Control display of class in records)
##                  +[no]cmd            (Control display of command line)
##                  +[no]comments       (Control display of comment lines)
##                  +[no]crypto         (Control display of cryptographic fields in records)
##                  +[no]defname        (Use search list (+[no]search))
##                  +[no]dnssec         (Request DNSSEC records)
##                  +domain=###         (Set default domainname)
##                  +[no]edns[=###]     (Set EDNS version) [0]
##                  +ednsflags=###      (Set EDNS flag bits)
##                  +[no]ednsnegotiation (Set EDNS version negotiation)
##                  +ednsopt=###[:value] (Send specified EDNS option)
##                  +noednsopt          (Clear list of +ednsopt options)
##                  +[no]expire         (Request time to expire)
##                  +[no]fail           (Don't try next server on SERVFAIL)
##                  +[no]identify       (ID responders in short answers)
##                  +[no]idnout         (convert IDN response)
##                  +[no]ignore         (Don't revert to TCP for TC responses.)
##                  +[no]keepopen       (Keep the TCP socket open between queries)
##                  +[no]multiline      (Print records in an expanded format)
##                  +ndots=###          (Set search NDOTS value)
##                  +[no]nsid           (Request Name Server ID)
##                  +[no]nssearch       (Search all authoritative nameservers)
##                  +[no]onesoa         (AXFR prints only one soa record)
##                  +[no]opcode=###     (Set the opcode of the request)
##                  +[no]qr             (Print question before sending)
##                  +[no]question       (Control display of question section)
##                  +[no]recurse        (Recursive mode)
##                  +retry=###          (Set number of UDP retries) [2]
##                  +[no]rrcomments     (Control display of per-record comments)
##                  +[no]search         (Set whether to use searchlist)
##                  +[no]short          (Display nothing except short
##                                       form of answer)
##                  +[no]showsearch     (Search with intermediate results)
##                  +[no]split=##       (Split hex/base64 fields into chunks)
##                  +[no]stats          (Control display of statistics)
##                  +subnet=addr        (Set edns-client-subnet option)
##                  +[no]tcp            (TCP mode (+[no]vc))
##                  +time=###           (Set query timeout) [5]
##                  +[no]trace          (Trace delegation down from root [+dnssec])
##                  +tries=###          (Set number of UDP attempts) [3]
##                  +[no]ttlid          (Control display of ttls in records)
##                  +[no]vc             (TCP mode (+[no]tcp))
##         global d-opts and servers (before host name) affect all queries.
##         local d-opts and servers (after host name) affect only that lookup.
##         -h                           (print help and exit)
##         -v                           (print version and exit)

To get the DNS records of r-project.org DNS we need to find the nameservers, which we can do via:

ns <- dig("+short", "NS", "@9.9.9.9", "r-project.org")
## ns1.wu-wien.ac.at.
## ns2.urbanek.info.
## ns1.urbanek.info.
## ns3.urbanek.info.
## ns4.urbanek.info.
## ns2.wu-wien.ac.at.

There are six of them (which IIRC is a few more than they had earlier this week). I wanted to see if any supported zone transfers. Here’s one way to do that:

stri_split_lines(ns$stdout, omit_empty = TRUE) %>% # split the response in stdout into lines
  flatten_chr() %>% # turn the list into a character vector
  map_df(~{ # make a data frame out of the following
    tibble(
      ns = .x, # the nameserver we are probing
      res = dig("+noall", "+answer", "AXFR", glue::glue("@{.x}"), "r-project.org", cat = FALSE) %>%  # the dig zone transfer request
        pluck("stdout") # we only want the `stdout` element of the {processx} return value
    )
  }) -> xdf

xdf
## # A tibble: 6 x 2
##   ns              res                                                  
##   <chr>           <chr>                                                
## 1 ns1.wu-wien.ac… "; Transfer failed.\n"                               
## 2 ns2.urbanek.in… "R-project.org.\t\t7200\tIN\tSOA\tns0.wu-wien.ac.at.…
## 3 ns1.urbanek.in… "R-project.org.\t\t7200\tIN\tSOA\tns0.wu-wien.ac.at.…
## 4 ns3.urbanek.in… "R-project.org.\t\t7200\tIN\tSOA\tns0.wu-wien.ac.at.…
## 5 ns4.urbanek.in… "R-project.org.\t\t7200\tIN\tSOA\tns0.wu-wien.ac.at.…
## 6 ns2.wu-wien.ac… "; Transfer failed.\n" 

(NOTE: You may not get things in the same order if you try this at home due to the way DNS queries and responses work.)

So, two servers did not accept our request but four did. Let’s see what a set of zone transfer records looks like:

cat(xdf[["res"]][[2]]) 
## R-project.org.    7200  IN  SOA ns0.wu-wien.ac.at. postmaster.wu-wien.ac.at. 2019040400 3600 1800 604800 3600
## R-project.org.    7200  IN  NS  ns1.urbanek.info.
## R-project.org.    7200  IN  NS  ns1.wu-wien.ac.at.
## R-project.org.    7200  IN  NS  ns2.urbanek.info.
## R-project.org.    7200  IN  NS  ns2.wu-wien.ac.at.
## R-project.org.    7200  IN  NS  ns3.urbanek.info.
## R-project.org.    7200  IN  NS  ns4.urbanek.info.
## R-project.org.    7200  IN  A 137.208.57.37
## R-project.org.    7200  IN  MX  5 mc1.ethz.ch.
## R-project.org.    7200  IN  MX  5 mc2.ethz.ch.
## R-project.org.    7200  IN  MX  5 mc3.ethz.ch.
## R-project.org.    7200  IN  MX  5 mc4.ethz.ch.
## R-project.org.    7200  IN  TXT "v=spf1 ip4:129.132.119.208/32 ~all"
## cran.at.R-project.org.  7200  IN  CNAME cran.wu-wien.ac.at.
## beta.R-project.org. 7200  IN  A 137.208.57.37
## bugs.R-project.org. 7200  IN  CNAME rbugs.urbanek.info.
## cran.ch.R-project.org.  7200  IN  CNAME cran.wu-wien.ac.at.
## cloud.R-project.org.  7200  IN  CNAME d3caqzu56oq2n9.cloudfront.net.
## cran.R-project.org. 7200  IN  CNAME cran.wu-wien.ac.at.
## ftp.cran.R-project.org. 7200  IN  CNAME cran.wu-wien.ac.at.
## www.cran.R-project.org. 7200  IN  CNAME cran.wu-wien.ac.at.
## cran-archive.R-project.org. 7200 IN CNAME cran.wu-wien.ac.at.
## developer.R-project.org. 7200 IN  CNAME rdevel.urbanek.info.
## cran.es.R-project.org.  7200  IN  A 137.208.57.37
## ess.R-project.org.  7200  IN  CNAME ess.math.ethz.ch.
## journal.R-project.org.  7200  IN  CNAME cran.wu-wien.ac.at.
## mac.R-project.org.  7200  IN  CNAME r.research.att.com.
## portal.R-project.org. 7200  IN  CNAME r-project.org.
## r-forge.R-project.org.  7200  IN  CNAME r-forge.wu-wien.ac.at.
## *.r-forge.R-project.org. 7200 IN  CNAME r-forge.wu-wien.ac.at.
## search.R-project.org. 7200  IN  CNAME finzi.psych.upenn.edu.
## svn.R-project.org.  7200  IN  CNAME svn-stat.math.ethz.ch.
## translation.R-project.org. 7200 IN  CNAME translation.r-project.kr.
## cran.uk.R-project.org.  7200  IN  CNAME cran.wu-wien.ac.at.
## cran.us.R-project.org.  7200  IN  A 137.208.57.37
## user2004.R-project.org. 7200  IN  CNAME r-project.org.
## useR2006.R-project.org. 7200  IN  CNAME r-project.org.
## user2007.R-project.org. 7200  IN  CNAME r-project.org.
## useR2008.R-project.org. 7200  IN  CNAME r-project.org.
## useR2009.R-project.org. 7200  IN  CNAME r-project.org.
## user2010.R-project.org. 7200  IN  CNAME r-project.org.
## useR2011.R-project.org. 7200  IN  CNAME r-project.org.
## useR2012.R-project.org. 7200  IN  CNAME r-project.org.
## useR2013.R-project.org. 7200  IN  CNAME r-project.org.
## user2014.R-project.org. 7200  IN  CNAME user2014.github.io.
## useR2015.R-project.org. 7200  IN  CNAME r-project.org.
## useR2016.R-project.org. 7200  IN  CNAME user2016.github.io.
## useR2017.R-project.org. 7200  IN  CNAME r-project.org.
## useR2018.R-project.org. 7200  IN  CNAME user-2018.netlify.com.
## useR2019.R-project.org. 7200  IN  A 5.135.185.16
## wiki.R-project.org. 7200  IN  CNAME cran.wu-wien.ac.at.
## win-builder.R-project.org. 7200 IN  A 129.217.207.166
## win-builder.R-project.org. 7200 IN  MX  0 rdevel.urbanek.info.
## www.R-project.org.  7200  IN  CNAME cran.wu-wien.ac.at.
## R-project.org.    7200  IN  SOA ns0.wu-wien.ac.at. postmaster.wu-wien.ac.at. 2019040400 3600 1800 604800 3600

That’s not pretty, but it’s wrangle-able. Let’s turn it into a data frame:

xdf[["res"]][[2]] %>% # get the response text
  stri_split_lines(omit_empty = TRUE) %>%  # split it into lines
  flatten_chr() %>%  # turn it into a character vector
  stri_split_regex("[[:space:]]+", n = 5, simplify = TRUE) %>% # split at whitespace, limiting to five fields
  as_tibble(.name_repair = "unique") %>% # make it a tibble
  set_names(c("host", "ttl", "class", "record_type", "value")) %>% # better colnames
  mutate(host = stri_trans_tolower(host)) %>% # case matters not in DNS names
  print(n=nrow(.)) # see our results
## # A tibble: 55 x 5
##    host                     ttl   class record_type value                                                              
##    <chr>                    <chr> <chr> <chr>       <chr>                                                              
##  1 r-project.org.           7200  IN    SOA         ns0.wu-wien.ac.at. postmaster.wu-wien.ac.at. 2019040400 3600 1800 …
##  2 r-project.org.           7200  IN    NS          ns1.urbanek.info.                                                  
##  3 r-project.org.           7200  IN    NS          ns1.wu-wien.ac.at.                                                 
##  4 r-project.org.           7200  IN    NS          ns2.urbanek.info.                                                  
##  5 r-project.org.           7200  IN    NS          ns2.wu-wien.ac.at.                                                 
##  6 r-project.org.           7200  IN    NS          ns3.urbanek.info.                                                  
##  7 r-project.org.           7200  IN    NS          ns4.urbanek.info.                                                  
##  8 r-project.org.           7200  IN    A           137.208.57.37                                                      
##  9 r-project.org.           7200  IN    MX          5 mc1.ethz.ch.                                                     
## 10 r-project.org.           7200  IN    MX          5 mc2.ethz.ch.                                                     
## 11 r-project.org.           7200  IN    MX          5 mc3.ethz.ch.                                                     
## 12 r-project.org.           7200  IN    MX          5 mc4.ethz.ch.                                                     
## 13 r-project.org.           7200  IN    TXT         "\"v=spf1 ip4:129.132.119.208/32 ~all\""                           
## 14 cran.at.r-project.org.   7200  IN    CNAME       cran.wu-wien.ac.at.                                                
## 15 beta.r-project.org.      7200  IN    A           137.208.57.37                                                      
## 16 bugs.r-project.org.      7200  IN    CNAME       rbugs.urbanek.info.                                                
## 17 cran.ch.r-project.org.   7200  IN    CNAME       cran.wu-wien.ac.at.                                                
## 18 cloud.r-project.org.     7200  IN    CNAME       d3caqzu56oq2n9.cloudfront.net.                                     
## 19 cran.r-project.org.      7200  IN    CNAME       cran.wu-wien.ac.at.                                                
## 20 ftp.cran.r-project.org.  7200  IN    CNAME       cran.wu-wien.ac.at.                                                
## 21 www.cran.r-project.org.  7200  IN    CNAME       cran.wu-wien.ac.at.                                                
## 22 cran-archive.r-project.… 7200  IN    CNAME       cran.wu-wien.ac.at.                                                
## 23 developer.r-project.org. 7200  IN    CNAME       rdevel.urbanek.info.                                               
## 24 cran.es.r-project.org.   7200  IN    A           137.208.57.37                                                      
## 25 ess.r-project.org.       7200  IN    CNAME       ess.math.ethz.ch.                                                  
## 26 journal.r-project.org.   7200  IN    CNAME       cran.wu-wien.ac.at.                                                
## 27 mac.r-project.org.       7200  IN    CNAME       r.research.att.com.                                                
## 28 portal.r-project.org.    7200  IN    CNAME       r-project.org.                                                     
## 29 r-forge.r-project.org.   7200  IN    CNAME       r-forge.wu-wien.ac.at.                                             
## 30 *.r-forge.r-project.org. 7200  IN    CNAME       r-forge.wu-wien.ac.at.                                             
## 31 search.r-project.org.    7200  IN    CNAME       finzi.psych.upenn.edu.                                             
## 32 svn.r-project.org.       7200  IN    CNAME       svn-stat.math.ethz.ch.                                             
## 33 translation.r-project.o… 7200  IN    CNAME       translation.r-project.kr.                                          
## 34 cran.uk.r-project.org.   7200  IN    CNAME       cran.wu-wien.ac.at.                                                
## 35 cran.us.r-project.org.   7200  IN    A           137.208.57.37                                                      
## 36 user2004.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 37 user2006.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 38 user2007.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 39 user2008.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 40 user2009.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 41 user2010.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 42 user2011.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 43 user2012.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 44 user2013.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 45 user2014.r-project.org.  7200  IN    CNAME       user2014.github.io.                                                
## 46 user2015.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 47 user2016.r-project.org.  7200  IN    CNAME       user2016.github.io.                                                
## 48 user2017.r-project.org.  7200  IN    CNAME       r-project.org.                                                     
## 49 user2018.r-project.org.  7200  IN    CNAME       user-2018.netlify.com.                                             
## 50 user2019.r-project.org.  7200  IN    A           5.135.185.16                                                       
## 51 wiki.r-project.org.      7200  IN    CNAME       cran.wu-wien.ac.at.                                                
## 52 win-builder.r-project.o… 7200  IN    A           129.217.207.166                                                    
## 53 win-builder.r-project.o… 7200  IN    MX          0 rdevel.urbanek.info.                                             
## 54 www.r-project.org.       7200  IN    CNAME       cran.wu-wien.ac.at.                                                
## 55 r-project.org.           7200  IN    SOA         ns0.wu-wien.ac.at. postmaster.wu-wien.ac.at. 2019040400 3600 1800

FIN

Zone transfers are a quick way to get all the DNS information for a site. As such, it isn’t generally recommended to allow zone transfers from just anyone (though trying to keep anything secret in public DNS is a path generally fraught with peril given how easy it is to brute-force record lookups). However, if r-project.org zone transfers stay generally open, then you can use this method to keep a local copy of r-project.org host info and make local /etc/hosts (or the Windows equivalent) entries when issues like the one this past week arise.

I can’t help but picture the 15-year-old — or 45-year-old — who’s never programmed before, and who reads a little bit about Ruby on some website and is intrigued and wants to try learning it.

Maybe they have a thing they want to do, or maybe they’ve always just been curious about programming and this seemed like a nice way to start.

Are we going to ask that person to figure out how to install Ruby? If there are one-click installers out there, are we going to ask them to figure out which one is actually reputable and safe?

Curiosity like this is one of the ways new developers are made. I worry that the less the Mac is tinkerable out-of-the-box, the fewer developers we’ll get.

Or: we’ll only get certain kinds of developers — the ones of the right age and background who can go get a CS degree.

This is a post about gluing some things together, and about the IndieWeb. It started, however, with a photo of a child on a bicycle.

I liked that photo, and I wanted to express that somehow in a public fashion.

I didn’t want to leave a comment on that blog post, as my aspirations were simply to express admiration for the photo and its subject; I was looking for something more “hej!”

It turns out that there’s IndieWeb for that.

One way of thinking about the IndieWeb is “all the plumbing of corporate social networks, without any of the corporate social networks required.”

In other words, in this case, “a like button for the web.”

Another way of thinking about the IndieWeb is to focus on the Indie: it’s a decentralized jam that allows us all to bring our own tools to the table, but to interoperate.

In my case the tools I needed to glue together are FreshRSS (the RSS feedreader where that bicycle photo originally caught my eye) and Drupal, which I use to write this blog.

Click the Star in FreshRSS

The way I decided to make this all work is to wire up “favouriting” a post in FreshRSS to sending a Webmention.

So I do this:

and I cause this to happen:

Pull The Favourites

My original approach was to try to code up a FreshRSS extension that would make this all happen in real time; I quickly decided that I didn’t want to have to grok a new MVC framework to make this happen, and that real time liking didn’t really need to happen.

I decided, instead, to simply extract favourites from FreshRSS on a regular schedule and to create new Drupal posts for new ones I encounter.

FreshRSS’s “entry” table makes this easy, as there’s a boolean field for each entry called is_favorite:

I only want to create a post in Drupal once, so I created an additional table, favourites_rss, to track those I’ve already processed. The result is that I can extract the details of new favourites by this bit of SQL:

SELECT en.id,title,link,date,website,name 
   from freshrss_peter_entry en,freshrss_peter_feed 
   where 
     (freshrss_peter_feed.id = en.id_feed) and 
     (is_favorite = 1) and 
     en.id not in (select fav.id from favourites_rss fav)

This returns me all the information I need about each favourite:

• The title of the post
• The link to the post
• The date of the post
• The title of the blog where the post lives
• The link to the blog

Post the Favourites

I created a new Drupal content type called Favourite with fields for each of these pieces of information:

With that in place I can programmatically add new favourites in Drupal like this, in PHP:

foreach($favs as $key => $row) {

	$node = new stdClass();

	$node->type = $nodetype;

	node_object_prepare($node);

	$node->language = 'und';
	$node->title = html_entity_decode($row['title']);
	$node->uid = 1;
	$node->status = 1; //(1 or 0): published or not
	$node->promote = 0; //(1 or 0): promoted to front page
	$node->comment = 0; // 0 = comments disabled, 1 = read only, 2 = read/write
	$node->field_feed_title[$node->language][0]['value'] = html_entity_decode($row['name']);
	$node->field_website[$node->language][0]['url'] = $row['website'];
	$node->field_website[$node->language][0]['title'] = html_entity_decode($row['name']);
	$node->field_link[$node->language][0]['url'] = $row['link'];
	$node->field_link[$node->language][0]['title'] = html_entity_decode($row['title']);

	if($node = node_submit($node)) { // Prepare node for saving
		$node->created = $row['date'];
		node_save($node);
		webmention_send($node->nid);
	}
}

The post that got created for the bicycle photo is here; if you look at the HTML of that post, you’ll see that the link includes the CSS class u-like-of:

Naar het park

The call to webmention_send() is the secret sauce that sends a Webmention to the original favourites post, a Webmention that signals “I like you” because of that CSS class.

Release the Favourites!

With all my favourites now safely tucked away in Drupal posts, with Webmentions sent to their hosts, I can also expose everything I’ve favourited to all-comers.

Drupal makes this easy using Views, which I used to create this Favourites page, that lists them all, in reverse chronological order; I also used Views to create an RSS feed of my favourites that you can plug into your RSS reader, should you like (and to allow the river of love to overflow its banks ever further).

Although it took some fiddling to make all this happens, now that the fiddling is fiddled, it just works: I click the star in FreshRSS and the favourite appears on this list, in this RSS feed, and, should the blog I favourited the link from support Webmentions, as a “like” on the original post.

Wenn ich irgendwelche Technik entsorge, behalte ich stets die Kabel und die Netzteile. Netzteile gehen kaputt und müssen ersetzt werden. Kabel haben nie die richtige Länge oder die richtige Farbe. Oder den falschen Stecker. In meiner ganzen Sammlung mit viel zu vielen Kabeln gibt es nur eins nicht: Ein kurzes weißes Stromkabel mit abgewinkeltem Schukostecker und einem Kleeblatt am anderen Ende. Samuel schickt mir jetzt eins - das ist sein Bild. Hat noch jemand eins in seiner Kabelkiste? Länge zwischen 50 und 70 cm.

Great piece by John Gruber on Jony Ive leaving Apple and the end of the Ive era at Apple:

I think Tim Cook is a great CEO and Jeff Williams is a great COO. But who’s in charge of product design now? There is no new chief design officer, which, really, is what Steve Jobs always was. From a product standpoint, the post-Jobs era at Apple has been the Jony Ive era, not the Tim Cook era. That’s not a knock on Tim Cook. To his credit, Tim Cook has never pretended to be a product guy, which is exactly the hubris that John Sculley succumbed to back in the early ’90s, leading to the Newton being launched far before it was ready and the Macintosh platform languishing.

My gut sense for years has been that Ive without Jobs has been like McCartney without Lennon. Or Lennon without McCartney — take whichever analogical pairing you prefer. My point here is only that the fruit of their collaborations were, seemingly magically, far greater than the sums of the duos’ talents and tastes.

Assuming Ive's exit has been planned for a while, it makes little sense to me that Apple's new design leaders (Hankey and Dye) are reporting to Jeff Williams, the company's chief operating officer – unless Apple has bigger plans for him in the near future. Then again, Gruber's point still holds: who's Apple's Chief Design Officer now?

→ Source: daringfireball.net

mkalus shared this story from mrjamesob on Twitter.

---

mkalus shared this story from Telepolis.

---