Good user support case study.

Topics:

• premium versus gratis support
• rude, abusive, or demanding users
• Discourse as a medium for handling support requests
• asking bad questions versus how to ask good questions

Don't do that.

Don't support this.

We can/should definitely go further and not even require a NodeJS.

Fix the URLs with this snippet (you can enter this directly into the JS console):

[ ...document.querySelectorAll("link") ].forEach((x) => {
  const good = "http://jonudell.net";
  const bad = "http://weblog.infoworld.com";
  if (x.textContent.startsWith(bad)) {
    if (x.parentNode.nodeName == "item") {
      let enclosure = x.parentNode.querySelector("enclosure");
      let url = enclosure.getAttribute("url");
      let fixed = fix(url);
      enclosure.setAttribute("url", fixed);
      let details = x.parentNode.querySelector("description");
      details.textContent = details.textContent.replace(url, fixed);

      let slugEnd = x.parentNode.querySelector("title").textContent.
        toLowerCase().
        replace(/[ /]/g, "-").
        replace(/[^a-z0-9-]/g, "");
      x.textContent = fix(x.textContent).
        replace(/\.html$/, "-" + slugEnd + ".html");
    } else {
      x.textContent = fix(x.textContent);
    }

    function fix(url) {
      return url.replace(bad, good);
    }
  }
})

In Firefox you can copy the result to your clipboard by following these steps:

1. From the page context menu, Select All
2. From the selection context menu, View Selection Source
3. In the source viewer, deselect the partial selection
4. Use Select All again to select the entire output
5. Copy the new selection

You should be able to dump this into a text file, copy it to a web server somewhere, and point your podcast app at its URL.

(The snippet itself should also work in Chrome but figuring out how to get the result into a file is an exercise left up to the reader.)

Should be "turning"

Should be "Its"

As it should be. (This should be the default. The NodeJS/NPM's Webpack addiction needs to be curbed—or confronted with a reality check.)

Vipercard as I understand it is more like a loose clone of HyperCard than a re-implementation. I think this page confuses Vipercard with HyperCard in multiple places.

The Internet Archive has done a lot of good work preserving HyperCard stacks, so you can experience them in your browser.

Need to add Newspeak and Hopscotch (Bracha, Bykov) to the catalog.

Need to add browser devtools (including sepsis and the original DOM Inspector) to the catalog.

should be "Snapshots"

Also available as PDF: http://bracha.org/jigsaw.ps

Also available as PDF: http://bracha.org/oopsla90.pdf

The implicitness that this "protocol" pushes into the implementing methods (seemingly odd return values) is undesirable. Make it explicit. The implementing methods should be able to do something like Validator.changeExpectation(this, this.onresource) (which in turn changes the target's ontokencontent handler).

No

Around 1:45:50

"Code is open source, but who cares? Nobody can understand it."

---

Yeah, and it's useful. Like: people do stuff with it, but as soon as you run into problems with a library, it's just like... looking into the code is a mess... There's just a huge number of problems with it.

"Not a valid community"

Near the end (@1:50:32):

My website is glench.com, and that's kind of my repository of everything I've ever made

Around 1:48:00

What if every library that you use had, like, some interactive documentation or interactive representation? [...] The author could maybe add annotations.

Non-shitty link: https://github.com/futureofcoding/futureofcoding.org/blob/main/episodes/052.md

The form action points to a server that is, as of this writing, not available.

Change the form action to point to http://harmonia.cs.berkeley.edu/harmonia/download/download-source.html, and the downloads are accessible.

Or, for completeness, here are the direct links:

• https://people.eecs.berkeley.edu/~harmonia/harmonia/download/harmonia-3.0.tar.gz
• https://people.eecs.berkeley.edu/~harmonia/harmonia/download/harmonia-3.0.zip
• https://people.eecs.berkeley.edu/~harmonia/harmonia/download/harmonia-3.0.tar.bz2

A casualty of the "consultant effect": under the guise of solving (existing) problems, generating even more complexity to justify the consultant–client relationship.

Compare this to algorithmic complexity; if a fancy algorithm is great for really large input sizes but comes at a cost of being somewhat worse for small input sizes, then it is a good idea to use the fancy algorithm when you might be dealing with large inputs. But if you are not dealing with large inputs, then using the fancy algorithm is a positively bad idea. It's a bad idea in the sense that it's not just a matter of YAGNI, but the fact that its characteristics for small inputs means that it actually makes the situation worse.

Same thing with NodeJS modules vs TC-39's native modules.

It bears repeating: NodeJS is a "vendor" for an incompatible fork of JS.

Good. Don't. The number of people who think this is a virtue is frightening. The rationale is usually not well-reasoned and whatever values they pretend to hold can almost always be shown to be hollow.

This is how I feel about most people's personal websites. Few people have homepages these days, but even for people who do, even fewer of those homes have anyone really living there. All their interesting stuff is going on on Twitter, GitHub, comments on message boards...

Really weird when this manifests as a bunch of people having really strong opinions about static site tech stacks and justifications for frontend tech that in practice they never use, because the content from any one of their profiles on the mainstream social networks outstrips their "home" page 100x to 1.

See also Lawrence Kesteloot's Java for Everything.

Here's what RG Arns, Crawford's advisor had to say about this topic wrt to Crawford's work in Arns's later article "The other transistor: early history of the metal-oxide-semiconductor field-effect transistor" for Engineering Science & Education Journal in 1998:

In 1991 Bret Crawford, in his physics MS thesis reported an experimental reconstruction of Lilienfeld's US Patent No. 1,900,018 using technology that was available in 1928, when the patent application was filed, and known to Lilienfeld. The basic idea of the device is shown in Fig. 3. Following the prescriptions in the patent, he produced working devices, but they were unstable and the results were difficult to reproduce. That was not surprising since the semiconductor, cuprous sulphide, is no match for modern single-crystal silicon. Crawford also found signs that Lilienfeld had actually built the devices he patented—and not merely thought about them—in the form of close agreement between phenomena described in the patent and his own observations.

Arns goes on to say:

In 1995 Joel Ross replicated the prescriptions of the same Lilienfeld patent. He was able to produce devices that remained stable for months. Although these devices showed the field effect, the transconductance was poor, apparently due to surface states. The power gains, however, were significant.

https://doi.org/10.1049/ESEJ%3A19980509

https://cacm.acm.org/magazines/2014/1/170862-actually-turing-did-not-invent-the-computer

This commenter is committing what I call "the timeless fallacy". It's not that far off from people who don't "get" evolution by natural selection—which is odd, because most of the people who commit the timeless fallacy in venues like this almost certainly do understand it.

This comment in 2010 highlights a difference between the world at large at that time (what I've called "the Shirky era") versus the behavior of users today

Why "broken"? (And why "dated" or "promising"? Tools for Thought is the name of a book. This is an online copy.)

"untrusted"?

Hypothesis should highlight this as one of its killer features: private "replies". More permanent than a draft.

(Unfortunately, a bug in the Hypothesis server means that this page is a 404 even for me—the person who wrote the reply, trying to view it while logged in...)

From Barack Obama's Dreams From My Father:

With our eyes closed, we muttered the same words, but in our hearts we each prayed to our own masters

Both module names and procedures are allowed to have non-leading digits.

bites? bytes

This should also be a null check, instead of falsy, else we can't read a hex-encoded NUL byte.

Tokenizer.readHex expects a 32-bit hex literal padded to 8 chars.

Should be 0xFFFF0000.

Unnecessary check.

Unused; this should be _glider.

From https://hypothes.is/a/Drc_4AtsEey4b3PQ82E8hQ:

The indentation here is supposed to be a tab, to match the output of ORTool. Same for the strings section.

The indentation here is supposed to be a tab, to match the output of ORTool. Same for the strings section.

At this time, this series does not appear in the search results on pbs.org.

From the Conclusion section in this paper:

Note: This paper is based on a keynote paper delivered by the authors at the WILUconference held in Guelph, Canada, in May 2005 (http://dis.shef.ac.uk/literacy/adelaide-webber-johnston.pdf)

Only by convention, for some server software.

To assemble 512 bytes? We can do better than that.

Nope.

Opportunity here to make things more "literate"; the transformation (with the representation of the result shown immediately following the place where this annotation is anchored) should be describable in language.

See also: the decision by every Web browser vendor's devtools team to make their viewers modal (in the 21st century!) and then add injury to injury making the monolithic implementation a "UI singleton"—so you don't even have the choice to open another instance!

Jekyll is an independent project from GitHub. GitHub by default supports Jekyll static sites because Jekyll was a popular static site generator in the Ruby community, and the GitHub people were Rubyists.

Agreed, but you can do even better by first making it available from your org's website. If there's any kind of fanbase at all, you can alert them to this development, so they can take make copies for safekeeping, taking a page from the LOCKSS strategy.

The Archive is great and doing a great service, but one should be wary of any assumption that they are an unlimited resource that will be around forever merely due to their best efforts, not to mention concentrating so much dependency on a single point of failure.

It's possible that the URI gives it away, but this is a local copy of the resource at https://colbyrussell.com/meta/pages.app.htm.

(~/projects/colbyrussell.com/main is the Git repository of all the Markdown sources and whatnot. It turns out that this file, though, is already in its file, publishable form and gets passed through much the same way e.g. a PNG or JPEG would be.)

"[...]of the document"

See < http://hdl.handle.net/10863/4821>/https://bia.unibz.it/discovery/fulldisplay/alma991005773520701241/39UBZ_INST:ResearchRepository

pp 17–18:

M. Van Swinden is the only person who has placed on record the observation of the dark lines, or threads, which connect the borders of the sun and moon, at the formation and dissolution of the annulus. His account is inserted in the first volume of the Memoirs of this Society (page 146), accompanied with drawings, which coincide almost exactly with those given by Mr. Baily.

It's tricky enough† to get a hold of Memoirs volume 1, but even after finding a copy, the only thing I can see in the vicinity of page 146 is Van Swinden's account, not the drawings.

† I couldn't find a full copy last time I tried in 2017, even though I spent a few hours reasoning at the time that surely I'd be able to turn something up...

Look here: http://profs.etsmtl.ca/mmcguffin/research/zigzag/

A non-PDFied HTML version is here: https://web.archive.org/web/20200830204849/https://journals.tdl.org/jodi/index.php/jodi/rt/printerFriendly/131/129

Unsure of the text encoding here. I'm forcing them to be interpreted as Unicode here, hence the appearance of the replacement character. My browser's default is to treat this document as "Central European (Windows)", but in that case, they appear as majuscule and miniscule S-cedilla characters (e.g. Şhypertextş).

By a reasonable guess, these are supposed to be open and close quotes. I've seen these appear in other TBL-authored documents from the same era.

Previously available under the URL https://corytheboyd.com/posts/2020-03-09.html.

Hypothes.is search: https://hypothes.is/search?q=url%3Ahttps%3A%2F%2Fcorytheboyd.com%2Fposts%2F2020-03-09.html

You're looking for http://web.archive.org/web/20051207225702/http://www.oreillynet.com/pub/wlg/7567.

This one: http://www.43folders.com/2005/08/17/life-inside-one-big-text-file

Archive available here: http://web.archive.org/web/20051207225702/http://www.oreillynet.com/pub/wlg/7567

Of "Dealers of Lightning", Alan Kay says:

Its flaws are too much "Heroes' Journey" and a very complex and confused jumping around timeline (I had trouble myself orienting in some of the spots). But it also has a lot of good stories, of which a reasonable number are "true enough".

He recommends Waldrop's "The Dream Machine".

https://news.ycombinator.com/item?id=22379275

You are looking for https://licsjournal.org/index.php/LiCS/article/view/794

This URL is eferenced from http://annettevee.com/.

Referenced from https://timbl.github.io/pad/

Huh?

Not only is this terribly drafted because it doesn't address the root issue of what was wrong with the MAI v. Peak decision, but it inadvertently ends up legalizing much of what would have otherwise been infringement for a class of traditional software piracy.

It doesn't specify for whom or limit the number of copies or relevant machines. So whereas before, you might be required to buy N copies of Microsoft Foobar, you are now permitted, legally, to purchase one copy, and then make another copy "as an essential step in the utilization of the computer program in conjunction with a machine", and then another, and then another...

In other words, if you needed 200 licenses before, for every CPU in your business, you can now get by spending 0.5% of your previous costs.

There's a lot of cruft here. Consider that while a project might have a prominently named file like "README" that is meant to be the first thing a wanderer encounters, the true first encounter is the file listing in the project source tree:

• build/
• config/
• src/
• .babelrc
• .dockerignore
• .editorconfig
• .gitignore
• .stylelintrc
• .travis.yml
• Dockerfile
• Gruntfile.js
• LICENSE
• Procfile
• README.md
• aldine.sublime-project
• aldine.sublime-workspace
• circle.yml
• package.json
• tsconfig.json
• tslint.json
• yarn.lock

Imagine a commit (or a pull request) with the summary "Remove cruft". Why might it be rejected? Let's get more specific.

There's a Dockerfile here. There's also a package.json. We can ask of each of these, "Why is this here?" The answer is, "Because someone found them useful." Consider, then, that here's a strong case for a contrib/ directory† for this project and where these things should be kept, ill-conceived tooling conventions notwithstanding.

† This link points to a particular blog post that explains the purpose of a contrib/ directory, but this is not an endorsement of Mr DeVault's other positions or demeanor. Ignore any stridence, arrogance, or other obnoxiousness that you might encounter in your pursuit to pull at any threads from that corner of the Web.

While nature doesn't, human systems have intentionality.

Ditch the perspective about the world that leads to thinking that everything is redstone.

See also: What colour are your bits?.

Referenced in DSHR blog post responding to Zittrain on Internet Rot

Really, this is an argument for self-containedness and not an argument against HTML and HTTP.

Granted, the Web doesn't handle compound documents very well (embedded graphics, unless they're SVG—and even then...). See https://blog.jonudell.net/2016/10/15/from-pdf-to-pwp-a-vision-for-compound-web-documents/.

Shouldn't the attribution info on this page ("admin", "July 26, 2021"...) be presented instead as Visual-Meta? Is Visual-Meta on the Web not important?

You're looking for https://www.youtube.com/watch?v=_cBGtvjaLM0

on the topic of interoperability one idea that I'm excited about is thinking about better ways to synchronize across existing cloud applications. I think there's a way in which if you're using one app and I'm using a different app and if we can establish a bridge between them, where let's say I'm editing a doc in google docs and your using Dropbox Paper or your preferred editor[...] that starts to create this more flexible feeling where the data's not locked in any individual app and it more kind of lives between the apps. And so one new project that I'm sort of embarking on now is trying to create tools that mediate that kind of synchronization across tools.

What if we used... files?

[nightmare involving a user-friendly bicycle]

There is a such thing as inappropriate use of abstraction.

For more on this https://news.ycombinator.com/item?id=27662074, and refer, for example, to NPM.

In software development itself, if you unleash a bunch of programmers on a problem and allow them to pursue their whims, you will observe that code tends toward bloat. This is not necessarily widely recognized. More broadly understood is the corollary to this that goes by the name Wirth's law, which states that software gets slower faster than hardware gets faster.

Outside of computer programming, there is a general awareness that organizations become less efficient with size. For the same reason that this happens, procedural bloat afflicts SOP just like code bloat happens with programmers.

There's a widespread belief that capitalism seeks out efficiency. With most organizations being capitalist enterprises, so the belief continues, they are an extension of this. You can see this show up in arguments about the gender pay gap. If we could cut costs just by hiring women to do the same job, they say, then we would. The veracity of the claims about the size of the pay gap notwithstanding, the claim that corporations would seize the opportunity to cut costs like this doesn't jibe with reality. Corporations are not observed to be a perfect extension of the law of capitalist efficiency. A corporation as an entity is not a perfectly rational actor operating in its own self interest, following both from the irrationality of the people who make it up and from instances of where they do behave rationally operating in their own individual self interests, counter to the organization's.

There is hardly ever a Taylor-like figure assigned to the problem of wrangling inefficiency.

Missing "Getting HTTP right" https://www.infoworld.com/article/2671157/getting-http-right.html.

Running the tests should be as easy as opening something in the browser.

How do we keep people from falling into the kind of rut that results in documentation like this?

I think it comes from an imposed milestone to document everything, so people end up phoning it in like this. In Graham Nelson's Narrascope talk (the one that was a followup to his broken promise that Inform 7 would be open sourced), you can see in his screenshots various passages filled with similar kinds of (frankly unhelpful) "prose".

Boooo...

Moved? https://tinlizzie.org/IA

See also, McCarthy's "ho, ho..." moment. https://hypothes.is/a/A_hmqoHZEeuEaEfF5uimPw

Antidotes:

• Carefully entrusting veto power to those who know the distinction between thoughtful caution as in the case of Chesterton's fence versus the kneejerk antidisestablishment response—similar to the case of the case of Rhesus ladders
• teaching the value of yes-anding over no-don'ting.

Is smoothsort stable? This quote seems to be saying it is, but elsewhere people say otherwise.

"A Constructive Approach to the Problem of Program Correctness."

Agree with the sentiment, disagree with the "remedy". (I realize it's not actually being proposed as such.)

colon-slash-slash was great in hindsight because there's sufficient entropy that if you want to do URL sniffing, you can get by with writing a really naive implementation that just looks for that sequence—you don't even have to worry about the "http" or "https" parts. In fact, I think it would be great if user agents came to grips with the dominance of HTTP and allowed links of the form starting with "://"—where the scheme name can be omitted entirely, and it is just presumed to be HTTP/HTTPS. (Use the same discovery method that browser address bars already use for working out whether "http" or "https" is the correct to go with.)

Huh?

This is really just making a strong case for the Web's notion of content negotiation, which receives pitifully little attention. The idea is that there should be some nugget comprising a cruft-free version of the resource's content itself, and it should be possible for you to specify, e.g., "No really, just elide all the JS and other accoutrements of the modern Web from the representation that you send me; basically, just give me the meat of it."

A thought by way of the Nelsonian school of hypertext:

Tim is quoted several times in this piece, which is to say that there is a larger body of content (say a recording of the interview, whether audio or a full transcript) from which this piece is borrowing snippets. Within the WWW worldview, that full record comprises a resource. Within the Nelsonian worldview of visible connections, at every place in the document where Tim is quoted, there should be an edge which the reader can traverse to reveal the unabridged record.

Not exactly an if-by-whisky, but (unrelated to the content of this article) worth serving as the launch pad for a series of examinations about why people feel compelled to couch their messages in this way—usually because the other "side" is using an appeal to emotion or appeal to reflex.

One thing I'd like forward-looking hypertext toolmakers to keep in mind is the ability for the tools to help people answer questions like "What led to legalanthology.ch hosting a copy of this document? Given a URL from one organization, is it possible to look at the graph of internal backlinks (let me focus narrowly on incoming edges originating from the same host)?"

This is where people with a tendency to participate in JS development frequently start to go off the rails. Exercise some restraint and tell that voice in your head that has been influenced by years in the community "no".

(PS: typeof checks need not and should not be written to use triple equals. Again, this is an example of where the dominant culture of modern JS development is a bad influence—pushing people towards doing things poorly. It's like radio and TV announcers who go out of their way to say "backslash" in URLs—"stop! you're going through extra effort just to get it wrong.")

Another broken link.

Broken link. See https://hypothes.is/a/JFM4yvuIEeuujaO6c6FwyA

You're looking for https://solidproject.org/self-hosting/css.

See also the sepsis inspector project.

or BYFOB.

With S4, we can do better.

How not to futureproof your work.

"Console application" is a Windowsism. Even the new Microsoft can't escape the strong pull of its legacy.

This URL is linked (in error) from the The Screening Room index.

You're looking for http://jonudell.net/udell/2006-10-01-the-screening-room-9-windows-live-writer.html

From Linus Torvalds Remembers the Days Before ‘Open Source’:

Torvalds counters that “I wouldn’t trust Lyle Ball’s recollection 100% about me… since my girlfriend-at-the-time (now wife) name was Tove, not Ute.”

Neat.

The incident described in this post provides a good case study for why the GitHub pull request process is not a good substitute for wikis.

Regarding missing 1993/1994 archives:

See https://lists.w3.org/Archives/Public/www-talk/2016JulAug/0005.html

See also https://www.w3.org/Proposal

Try https://www.w3.org/Administration/HTandCERN.txt

You are (probably?) looking for https://www.w3.org/Addressing/URL/Overview.html

Referenced from http://info.cern.ch/hypertext/WWW/Bibliography.html.

This page is now a 404, and the Wayback Machine (whose first access attempt is dated 2013) doesn't have a copy. (It was already a 404 by then.)