ttj - Tom Jones

All the departure screens in the airport are showing this: According to staff, it's been broken since friday (today is monday). The issue was described as being "Stateside", by which they meant it was the responsibility of the States of Guernsey. In other news, Aurigny are unable to easily bump you onto an earlier flight if you turn up early and there's space, something BA used to be good at, back when they were an airline.

There are two types of bullshit that commonly go together.  I say bullshit rather than mistake, or fallacy, because we are in the realm of narrative-pushing.  The first type of bullshit is confusing the total value of a thing with its evamp, or extrapolated total value at marginal prices [0].  This is a form of innumeracy.  The second type of bullshit is fitting narratives to price fluctuations.  Doing so makes unfounded claims about what the bullshitter knows about the causality involved, and in any case is often applied to noise rather than real moves, allowing the bullshit narrator to select whether the item is "up on" or "down on" the event they are bullshitting about. [0] < https://wibblement.blogspot.com/2021/11/evamp-market-cap-evamp-extrapolated.html > The particular case here came to my attention via this tweet:     in which they claim "[..] Ronaldo's removal of [bottles] coincided with a $4bn fall in the company's value".   But, on

Blogger is not a tool that can be ameliorated into something good; it would be better to start from scratch.  One quirk: to add a picture for illustration, a lot of back and forth is required between the web page and their service.  Since picking a file, or picture, is a standard form control, the amount of back-and-forth needed is zero.  Unless the internet connection is fast, consistent, and with no packet loss, blogger fails to bring up the over-complicated dialog that is the next step in selecting your picture.  This is silly.  And the connection here is not that bad -- it's been fine, even fast, for browsing a variety of other web sites. This does bring up a serious point, which is that web (and other) software should explicitly state the connection parameters it targets, and it should then be tested with those parameters, instead of just over the LAN from the office next door.  You don't have to actually go on the train, you should be able to set up access through some ki

 " mount" , as a utility, is not defined by POSIX.  I don't know if it has been defined by a previous version of POSIX, or what versions of Unix, back when versions of Unix were specified, had it defined.  As a traditional Unix user, I would like it to tell me what  devices are mounted where in the system's filesystem namespace, that thing that starts at "/" and goes from there. For the currently-running system, then, this would look something like: /dev/mapper/sda2_crypt on / type ext4 (rw,relatime,errors=remount-ro) My main quibble with this is that the device Debian are calling "sda2_crypt" is actually the unencrypted map of the underlying device, not the encrypted one as the name suggests.  The FDE, the way this relates to block-device naming, this "mapper" thing, and how it's manipulated, from the user perspective, could all do with some design, and design is not something linux does, but I would still be more or less happy with

  I've just bought a server with 1.5TB of RAM.  At 1-2MB per process, that would be running a million processes.  That might not make sense with only a few dozen cores, but running more than 2^15, or ~30k, processes, which, as far as I can tell is the current limit in Linux, must be useful sometimes. According to a part of something calling itself " The Open Group Base Specifications Issue 7" [0] (I'm not sure, but this may be stuff common to both the C and POSIX standards)   : blksize_t , pid_t , and ssize_t shall be signed integer types. The relevant one is pid_t.  It doesn't place a constraint on the size.  What would break if we had a system with 32 or 64 bit signed pid_t ?  (Is it common for client code to make wrong assumptions about this?) [0] https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/sys_types.h.html What error do we currently get if the system runs out of pids? My basic idea is that the multi-user nature of the existing Unix design (

There should be a direct-link protocol.  Pairing (or equivalent) a pair of devices once should result in an authenticated link over any mutually-supported underlying link such as bluetooth, usb or wifi.  The same services should be available over any of these links (where physically possible) whether accessing files, transferring photos, sharing internet, and so on. Instead, there is currently no unity across the links.  Bluetooth pairing is one thing, wifi might be used to share via a hotspot with its own WPA password, USB asks you on the screen if you trust your laptop (each time) and then presents files over MTP, which is okay for photos but horrible for general filesystem hierarchies.  There doesn't even seem to be a sockety fs protocol over USB: it's either block device, which is the wrong level, or MTP, which is unsuitable.  It should look like nfs or 9p, adapted somewhat for the purpose, stuck in its own USB profile.  Usually the best way to transfer a picture from your

  In-thumbnail text is "Putin goes for Poland".  This was already looking dubious when it was in my feed on the evening of Nov 15th.  Just noting it briefly as this one's clear cut, and illustrates how full of shit The Times is -- at least as bad as Telegraph, Guardian, BBC etc.  Was surprised to see it there unchanged now. BTW I need to start running my "yt-dlp" jobs with " --write-thumbnail ", as the thumbnails often contain critical information, including textual information larger than the actual title text and acting as the video title.  And WTF, blogger editor's "default font" doesn't work (stayed as fixed-width), and I don't know which (if any) of the fonts in the list is the default one.

Previously: " Amazon: amatuers " [0] [0] < https://wibblement.blogspot.com/2022/11/amazon-amateurs.html > The user journey ended in a dead end anyway: It says "see below", but there is no further information below; in particular no reason for the "problem" is given anywhere on the page.  Since the web is now "mature", we have probably reached an equilibrium for actual links vs vague phrases like "our user manual", "below", "section 3 of our customer portal", and so on.  Of course, the (non) hypertext issue is secondary to the absence of information on why the purchase can not proceed. Apparently Jeff Bezos always asks for a "root-cause analysis" to be performed, every time an issue is identified.  Even tho Amazon isn't paying me for this, yet, I've performed some "root-cause analyses" for these issues.  The root cause of both these issues, as well as the root cause of Amazon's au

No one's pretending google has ever been of use for anything scholarly, or is structurally capable of becoming so, but I was tickled by this, brought up when I typed in "Odyssey" to check my spelling:

Just because Amazon is marginally less rubbish than other shops, does not make them any good.  They are going further and further down the path of being obstinately unable to authenticate their own customers, forcing re-"authentication" after re-"authentication", insulting time-wasting crap as displacement for being unable to authenticate.  Yesterday I had to identify cupcakes for bitcoin exchange Bitstamp, and today I get the following from Amazon: I was already logged in to my Amazon account.  I went to buy something, and suddenly I was not logged in after all, I had to type my password.  After putting in all the work to find the various punctuation characters on the android keyboard, my next task was: type in my password again.  This wasn't some sort of glitch, it explicitly said I had to type it "again".  I was faster the second time, knowing where the tilde lived.  Next, they forced me to solve a captcha, screaming "we are incompetent at auth

Take a pandemic nostalgia trip on a German train, where FFP2 masks are mandatory in November 2022.

  I can't be arsed to rant about it, other than to note that the entire edifice of shitfrastructure does not work.

A local heating firm just came to service the boiler.  This is part of the rented house, so I don't seem to have a choice in supplier.  Having finished the job, they put a blank screen in front of me, and asked me to sign.  The optimistic interpretation of this would be "signing a blank piece of paper means nothing; therefore signing this blank screen means nothing.  My signature is not itself confidential, and by providing it here I do not commit to anything.  At most, the implied meaning of signing a blank sheet for someone who's just done a job in the house is "they visited". Being conservative about what I sign, I asked, in very bad German, what am I signing here.  At this point he did something on the touchscreen, and there was a picture of an invoice on a sheet of A4 paper.  He continued to hold the device facing himself, and I was unable to read it upside-down and at a large angle and held by another person.  But the decision was already made (and really w

  Powering (or charging) a thinkpad used to involve a thick three-core mains cable with plug, a large transformer unit, and a DC lead integrated with the DC jack for the laptop end.  The full assembly above comes in just short of 500g.  It's always been 20V. It was possible to save some travel weight by making your own very short mains leads, splicing the DC lead, buying an alternative transformer, and so on, but not by much.  I even saved some weight by importing A Merkin lenovo transformers, which are perhaps class II appliances, but in any case had a two-connector figure-8 (C8) connector, only requiring a thinner 2-core mains lead.  But the transformer was always big. On their way to USB-C PD, thinkpads had a variety of other DC inputs: the square one, the thin round one.  Even when thinkpads, for example the X1 Carbons, started to have USB-C PD inputs more than 5 years ago, lenovo shipped them with bulky, heavy power supplies, with the transformer still in its own black box, an

 As I approach Paris, Interrail has entered a new and exciting failure mode: I appear to be logged in, and a search for the route even brought up some results.  I click on "Book Reservation".  I get to step 2/2, which sounds promising: One the one hand, this shows some of my details, providing further assurance that I'm logged in.  On the other hand, given that I only have one "trip" on my Interrail account, why does Interrail obsessively make me remind them which one of my one "trip"s I want to do things with, including every time I want to do anything with the app, including show my pass to a ticket inspector?  But more importantly on that other hand, clicking "Done" here does not go from step 2/2 to some kind of let's do this payment screen, but instead goes back to a screen just like the top screenshot of this post, which they entitle "Book Reservations", and which one might dub "step 1/2" to emphasise the incorrec

Repeatability is a perennial software engineering concern, so it may worry the team at interrail that their web site, in failing to sell seat reservations, does so in a slightly different way on different browsers. This morning, on chrome, we have this: The button has not been pressed.  It is never activated and says "Please wait", and is permanently in a faded state indicating non-activeness. On firefox, and with encouraging consistency, we still have: The "Please wait" on this was also never working. But, I should warn you, this is no time for complacency.  On the Android browser (also "chrome"), the login worked, and it was not until a subsequent stage that it became impossible to complete the reservation: In this case, the search button has been hit, but the results never come.  Plenty of effort has been put into the "user experience" that happens while the results don't come.  There are whizzy things going around, and gray templates, jus

 Could these items be in any kind of defined order? : UK - Guernsey UK - France UK - Jersey Guernsey - Jersey Jersey - Guernsey Sark - Jersey Jersey - Sark Guernsey - France Jersey - France Guernsey - UK Jersey - UK France - Jersey France - Guernsey France - UK It can't be by total volume of traffice, because then Guernsey - UK would be up with UK - Guernsey.  It's not some kind of "left item first, then right item", because the middle lot are all mixed up. Could it be "put UK-originated ones at the top, and the rest however they fall out of a hash table"?

Interrail login is now failing in a different way, with red writing "Something went wrong": previously: < https://wibblement.blogspot.com/2022/11/interrail-unable-to-sell-seat. html>

For some time it's been impossible to get original picture files back out of trello. Now, I can't find any way to get this pdf back out, to print.

 failed payment:

 The appeal of interrail is the prospect of being able to use railways without having to deal with unmanned stations, non-working ticket web sites.  The offline app works well enough to add journeys reasonably quickly.  But, in some of Europe's more backwards regions, "seat reservations" are mandatory.  The extremely backwards dens of indolence, such as France, seem so impressed with the smell of their own high-speed routes, they insist on them being treated specially.  Buying a ticket, sorry, reservation, on the spot, cost me 70 EUR(2022) last time (to someone reading from the future, this is several hours' work at median wage).  The interrail app says it's 12 EUR if booked online.  So I went to their web site to log in. Initially, the log in appeared to work, but then it went on a redirect frenzy, from login page to some other page and back and forth a lot, resulting in nothing working.  I closed the browser and started it again, and now I get: The "please

 Condor Ferries previously failed to sell ferry tickets due to a stuck date [0]. [0] < https://wibblement.blogspot.com/2022/11/condor-ferries-unable-to-sell-ferry.html > I tried it with my other browser.  It was still stuck, but on a different date, even further in the past:  Once again, it brings up a grayed-out month that isn't changeable: These are dates that I've used the service, so I guess something is going very wrong at their end involving a cookie.   So you get one shot at using the service per browser instance, or clear out cookies.

I went to Starlink's web site to buy their service.  It said "YOUR SESSION HAS EXPIRED", saying 15 minutes, but I had only just loaded the page. I went back to their front page and started again, but it failed in the same way again.  Like every other "e" "commerce" "vendor", Starlink is unable to cobble together a web site capable of the most basic vending.

I went to condorferries.co.uk to try and buy a ferry ticket.  The date was set to a date almost three months in the past: It was impossible to edit the date as text.  Clicking on it brought up a date selector -- a very bad date selector, which did not allow selection of a date: Consistent with the field's default value, the month shown is August.  All the days are grayed out -- presumably because they're in the past.  There is no visible way to change the month.  I hovered around where the arrows might be for change-month, if they existed, but nothing. This is the state of the web.  Since 1997, the probability of successfully completing any given transaction using the web has monotonically decreased from year to year, and now stands at almost zero.