161 thoughts on “It was the best of Memes, It was The worst of Memes

  1. I admit a mild disappointment. But I did not en-meme it myself, and the screen-grab did IDENT (such as is it is) the origin, so maybe it’s as well. For those wondering, text follows:

    The “anti-communists” are pro-forced-collectivism.
    The “anti-fascists” march in lock goosestep and commit thuggery.
    The “anti-sexists” are fine with men dominating women’s competitions.
    The “anti-racists” want to segregate people by color or ethnicity.

    It’s enough to give a fellow a case of ‘anti-pathy’ about all their supposed causes.

    1. Guy in the machine shop has a shirt that reads: “In case of emergency, my blood type is Mountain Dew”

      maybe that will work for you

    2. Caffeine reduces sinus swelling, and therefore makes it easier for sinus sufferers to go to sleep. Many’s the time that I’ve drifted off to sleep after sucking down caffeine and Tylenol.

      It’s also useful to normies, because you can power nap and then be woken up quickly by the caffeine.

  2. Seeing the last meme, I thought of the programmer’s nightmare.

    “Who wrote this shitty program! Oh Sh*t, I did twenty years ago.” 😆

      1. If I’m writing it for myself, of course I comment it. If I’m writing it under deadline for some minimum wage paying schmuck, he can add the comments after I resign.

    1. Always Comment your code, because the poor bastard who has to figure out what you were thinking might be you.

      1. “COBOL Code is Self-Commenting.”

        COBOL Programmer: “Hold my beer.” [Crazy Grin]

        1. “COBOL is organized code.”

          ANY COBOL program: “Hold my beer. Untangle this code knot. It is self commenting.”

          COBOL programmer. “Ahhhhhhhhhhhhhhhhh” breath “Ahhhhhhhhhhhhhh” breath “Who wrote this?” Pause. Either: “That programmer is lonnnnnng gone.” or “Crap. I did.”

            1. No thank you.

              I’m done programming. (30+ years professionally.) I swore off COBOL early (-ish).

              Something 41 years ago I’d never have thought I’d have said. (Intro to Computers. This time on Apple IIe’s.)

              OTOH 48 years ago I’d had sworn I wouldn’t ever have anything to do with programming again too. (Computer 101, Basic, Teletype. My typing was less than stellar. It’s improved somewhat, as long as I can correct as I go. Teletypes, not so much.)

              What can I say. I’m complicated.

              1. The Reader also hated teletypes with a passion – especially when he had to use it to punch Fortran source code to the teletype’s paper tape punch.

                1. You dealt with paper tape computer input?

                  I guess that I’m lucky that I only dealt with card input (for source code and computer input). 😉

                  1. To change a single line of Fortran or Assembler on the RF test equipment I worked out right out of college the process was as follows.
                    1. Type the source code on the teletype – didn’t matter whether it was one line or all new code – you had to type it all.
                    2. Load pass one of the Fortran compiler using the high speed (giggles) paper tape reader.
                    4. Load the paper tape of the source the same way.
                    5. Pass one of the compiler generates an ‘intermediate binary’ from the tape punch assuming there were no errors of any kind (HP’s term – the Reader never did figure out what was in it)
                    6. Load pass 2 of the Fortran compiler.
                    7. Follow it with the ‘intermediate binary’ tape
                    8 Pass 2 of the compiler generated an actual relocatable binary on the tape punch – except when it sent a cryptic message to the teletype instead.
                    9. Load the binary loader through the paper tape reader.
                    10. Load your program through the paper tape reader.
                    11. Load the library relocatable binaries through the paper tape reader.
                    12. The loader generated an absolute binary through the paper tape punch.
                    13 Load the absolute binary through paper tape reader.
                    14. Set the next address on the computer front switch register to the absolute starting address of the program.
                    15. Press run and pray.
                    Debugging was done by running your relocatable binary though a reverse assembler and looking at the assembly equivalent of your Fortran code.

                    Eventually steps 1-8 got replaced with an off line workstation that had a CRT terminal, text editor and compatible Fortran compiler and Assembler.

                    1. At least with punched cards, you could just change just one card and stick it back in the deck.

                      It was “interesting” to add new cards to the deck in the proper place.

                      And it was extremely “interesting” if some idiot dropped the card deck on the floor and “you” had the job of getting the cards back in order.

                      So CRT terminal, text editor and compliers were gifts from God! [Grin]

                      And yes, I’d hate to have dealt with paper-tape.

                    2. When I took COBOL I and II, as long as you numbered your cards (and there was a switch on the card punch machines to do that for you), you could hand the deck to the computer folks, and they could run it through a machine to rearrange it. It took time, which is why you got 1 free sort per quarter. After that it was $25. Most people got the hint and flipped the switch…..

                    3. Chuckle Chuckle

                      Punching numbers for the cards was what I was taught to do and when writing new programs it was great as long as long as you left plenty of room for inserting new cards (with numbers) for possible changes.

                      And that was the big problem for maintenance on existing programs. Did the last programmer leave enough “gaps” in the numbering for your changes? 😀

                    4. We were taught to number by 10s, then fill in the odd numbers between, then the evens, and I don’t remember needing to insert more than that.

                      Of course, by the time I graduated, everything was on screens.

                    5. So CRT terminal, text editor and compliers were gifts from God! [Grin]
                      ……………..

                      Amen

                      Although it did not stop someone from putting together OOP to give the one who came behind them a major headache.

                      I’ve mentioned “Universal Program Generator” (for DOS Falcon) (UPG) by Percon before. Some of you even said you’ve used it. There were 5 pieces, originally:

                      Main Program
                      4 sub-DLL’s for the four pieces that were actually used.

                      Sounds good. Right? Until the realization sets in that the Main Program had to compile, then compile the 4 DLL’s, then recompile the main program. Then maybe it would work. Because the core object definitions the DLL’s needed were in the main program. Every time an interface changed the above sequence had to be done. It never cooperated. Note, this was the originator’s attempt to make Visual Basic OO. It isn’t. Trust me.

                      The C code changes were simple in comparison, whether it was generated based on user creations or were part of it’s core library.

                      Third update (new hardware to support), marketing came to me and asked for the changes to be done in 9 months. I told them no. Minimum 18 months based on the list. Marketing went whinning to TPTB. Managers came to me asking why. Told them. Also told them if I would be allowed to upgrade, still not really OOP structured, but “better” which would force the reorganization properly, then I could may get the changes in 12 months. Which what the hardware side was pushing for. The whole “engineering said, marketing set the release date” premise. Six months to fix the organization, make sure hadn’t broken anything, actual changes 6 months. Had the reorganization done in 3 months, 3 months to make the changes (oops, OTOH made the 9 month deadline 😉 ). That was including how to get multi-byte characters saved and retrieved out of DOS text with standard US setting (doable, but wow is it buried deep in the MS C++ code library). That took me 6 weeks. (The books show “this should work”. Nope. Dig into the actual library. Will this function work. How do I get my code to recognize this function. Nope that didn’t work. Dig deeper. It has now been almost 25 years. Forget how deep I had to go. But essentially find the functions that use the byte code to be passed in. Easy. Once the correct functions were found.) Bigger pay off? Future upgrades, with even larger requirements, took 3 to 4 months. Way less than the hardware changes required.

                      Good news/Bad news. Made changes look too easy to do. Meant when forced cuts hit, I got on the chopping block about 3 months earlier that might have otherwise. OTOH another 3 months before getting desperate, might meant I missed out on applying for the job I did get. I hated applying to ads that did not list the company. This one didn’t. Desperate. And, yes, they did call me for help after I got cut, a few times. Replies: “I don’t know. I’m not there.” Yes, easy for me. Not so much to walk someone else through the code, over the phone, without being there (and not getting paid). I figured it out, without help, without documentation, I did document what I did and why. All the types of code used: Visual Basic, C++, and C. (I learned that lesson a long time before.)

                    1. Ah, you missed out on a lot of fun, for various values of masochism. I had to program with cards in HS and BS, ending around ’73.

                      Didn’t do any programming my first couple years ad Defunct Semiconductor, but the next job was on CRT and proprietary tape cassettes. Then was HP (for over 20 years, so things evolved). Started writing in more-or-less machine code for the Fairchild 5000c tester. Lines like

                      100045100000: 0105007A0000: 211010000000: 2A000001000:
                      (Connect supplies 4,5,and 1 to pins 3 4 and 5)
                      (Supply #1 has 5V with a 1A clamp. Goes on from there.)
                      (Going from distant memory; pardon the typos. 🙂 )

                      The techs told me that these instructions moved verbatim to the instruments. Easy to debug, hard to program. Back then, they were easy to remember.

                      At first the program were written on teletype and stored to 1/2″ mag tape. (Datalogs also went to mag tape for reduction/analysis on the big mini, an HP1000. A few years later, the in-department programmers persuaded the 5000cs to talk to the HP1000 for download and datalogs. This got us to programming on CRTs (finally!), though the coding/language stayed the same.

                      Somewhere in the early 90s, we bought some testers from a now-defunct company (died in the Dot-com/2001-2 recession) that used C. Good bye programming on the bare metal. My last job was consulting with the tester company as they were developing RF test capability. They died shortly after we finished the project. (Not cause and effect 🙂 –the CEO succumbed to Edifice Complex and the really nice new headquarters/manufacturing/development building combined with sales falloff to collapse things.)

                    2. I started the new career with CRT, keyboards, and regular sized floppies. First jobs after AA degree had floppies and small hard drives (System 36 mini, with IBM workstations). First job after second bachelors, all the IBM’s workstations had very small hard drives and were working with hard floppy disks. The *Xenix machine had 300 MB hard drive, and had moved to tape backup/restore. Interestingly enough still had old data stored on 8″ floppies.

                      ((*)) Interestingly enough had to upgrade (late ’95 to Pentium because of catastrophic hardware failure. That was fun. 1) Could not swap A, hard floppy) and B old floppy drives, because Pentium configuration would not allow that. Needed to swap to install Xenix OS. Called Xenix to get smaller hard floppies. Xenix won’t run on Pentiums. Got Unix instead. Meant getting COBOL for Unix. Only that type of change required 2 year testing cycle in conjunction with continuing the old system. Touching the data file with Unix compiled code (no changes in code required) changed the data files. So made sure had two sets of the data files, two sets of programs, one set for Xenix, one set for Unix, in addition we were working to move everything to SQL to work with Unix GIS system (didn’t get very far, but we had it designed, and were working on it). Turns out the Xenix programs would run on the new Unix OS, but couldn’t compile the programs on the Unix box. Luckily still had an old Xenix system installed elsewhere (duplicate system, allowed that site to run faster reports). So edit programs on the Unix system, send code to remote system dial up, compile on remote system, retrieve compiled program, test Xenix compile. Compile code to Unix program, test Unix compile, compare results. Meanwhile officially the Xenix programs were still the “official” programs. Four months later the shutdown was announced. That is the state the system was handed off in to the two partners that bought the timberland and accompanying assets (except most the main buildings and people). OTOH they had the option to cut to just using the Unix programs, without the two year testing cycle. Program was already written and tested to convert the COBOL data set to database that allowed easy importing to SQL. Or better yet, swap out the intermediate database code use to convert to SQL (didn’t have that library, yet). Also had the programs written in Unix to write the reports against the temp database structure, and tested. Again, all they needed was the SQL COBOL library interface, and to swap out any minor differences. Difference? Me, myself, and I, were constrained by lack of resources. The two companies each, had IT programming (just programming, not hardware too) departments of a minimum of 5 (which is why my managers didn’t have any luck convincing either party to hire me on. They each had more than enough programmers.) Other than this was my dream job (Forestry AND Computers), fact was my salary increased by a factor of 1.5x, about 2 weeks after severance and unemployment ended. But still. (We won’t discuss what happened salary wise 8 years later. Never been happy. OTOH better, at the time, than $0, and watching savings drain.)

              1. The Reader is surprised that D&D doesn’t have its own programming language. After all someone had time to create Klingon and Shakespeare.

                  1. Wow! Signetics made a microprocessor in the mid-’70s that had a very limited number of commands (all in a 40 pin(!) package). I had a sample, but my sense of sanity told me to not mess with it.

                    1. And I thought Brainfuck was a ludicrously ugly language. Someone needs to get these esoteric language language creators a job (like busting stones or laying brick, keep them the heck away from computers). Next thing you know the SOBs will be working on AI. Ooops, likely too late, behold Gemini.

                    2. <blockquote>Ooops, likely too late, behold Gemini.</blockquote>

                      Does the rise of Gemini mean that the Obfuscated C Contest has run its course? Asking for a friend…

          1. Relieved smile. My sole COBOL program came at the end of the one-semester high school programming class. Didn’t get it to compile (nor did most(all?) of my classmates) before the end of classes. Shiny new NCR in 1970; our HS was the first customer in that metro area. The vendor analyst/software whisperer was a fixture.

            Fortran (both II and IV) were bad enough. I can read C programs (even the ones I wrote 🙂 ), but the little I need to do now is in Perl or simply a shell script.

          2. You mean like the 64 lines of consecutive nested IF statements that decoded the UPSI switches into a single variable? That after ten years suddenly stopped working, thus crashing a critical program?

            I solved that one by inserting a GOTO that bypassed the whole mess and pointed to a new line that assigned a constant to that variable. Even tho I was the most junior programmer, nobody questioned why it was working again. 😉

            1. Goes to acronym-decoder for UPSI. Sees some that imply power systems, and one that I really hope doesn’t apply. (Unprotected Sexual Intercourse.)

              Switches? 🙂

              1. UPSI (User Program Switch Indicators) COBOL only – IBM

                My first programming job was in a banking environment, supporting COBOL programs running on a Burroughs computer.

                1. The NCR I learned on in high school replaced a Burroughs with all of 8K of magnetic core memory. 4 tape drives, and it managed to do what was necessary for a large high school (approx 4000 students). It was worked pretty hard.

        2. And then there’s Forth…
          It’s not quite ‘WRITE ONLY’ but it’s mighty close in the wrong hands. And most hands are so very wrong.

          1. … and APL.

            What’s the doggerel?

            “There are two things
            A man must do
            Before his life is done:

            Write two lines of APL
            And make the suckers run!”

            Tried to take an ‘informal’ lunch-time APL class in grad school. Gave up after one session.

            1. CS101 was required in engineering at U of Redacted, and was used as the big flunkout course. All but the last program set was in Fortran IV (on an IBM 360 mainframe), but the last one was in APL. (Had another class later on that had to do a project in Fortran II. Eventually TPTB cut the funding for that ancient computer.)

              I don’t recall any of the language. I can read Fortran, though I haven’t had to write any in a half century or so, but APL is lost in the mists of time. Whew!

              1. I can read Fortran, though I haven’t had to write any in a half century
                ………………..

                Not quite as long. But same.

                Big weed out coarse at U of (redacted) involved using Schema. I got to bypass that. I can read it, even without taking the coarse, because of all the other tools I’ve used. Never wrote in it. The only two requirements from undergrad CS I had to full fill were math (to be fair only had one term of the requirements, by then over 10 year prior, and it had not stuck, still don’t care for first term calculus), and a minor. Rolls eyes. Chose “forestry” as a minor. CS department clearing house staff came back “forestry isn’t offered at U of (redacted).” Me: “Already have a bachelors of Forest Management.” Department head: “Minor choice signed off on.”

          2. Had a passing acquaintance with Forth (grad school instructor had us do a something that used Forth, and I looked at the code). Didn’t get into the gory details (somehow I had a 3.5″ floppy with the language) but the Reverse Polish Notation system didn’t look that strange. OTOH, I’d been using HP calculators for over a dozen years at that time–now more like 50, so the mindset was set.

            1. The Reader remembers that in the late 70’s HP brought out the 9825 ‘desktop calculator’. It could be used to control instruments across the then HP-IB bus and was programmed in a (massive) extension of RPN called HPL that was Forth in all but name. We had them around for several years before they were overtaken by the 9835 series and HP Basic.

              1. We had a digitizing station (part of the transition from drawings to pure CAD for semiconductor designs) that used a 9825. Interesting machine, but not easy to deal with.

                We got 9836s for various things. Some of them ran under the Pascal P-system, though they could be rebooted to run Rocky Mountain Basic. One of the mask designers broke for lunch (hadn’t saved the file, either) and a co-worker (engineer) rebooted the machine to play games. Much tears, and I made a hint to the guilty engineer that I’d prefer not to rip his lungs out, but it was an option in case of a repeat performance. The mask designer had her own punishment–had to redo the morning’s work…

                I had a Heath H8 that ran the Pascal P-system. Didn’t care for the language, nor the OS. Eventually bought CP/M and liked it for the 8080.

                1. The first lesson every aspiring programmer needs to learn: Jesus saves! Hell is having to retype your code. Again.

                    1. I still remember Jerry Pournelle’s column in Byte where he was dancing the Happy Dance over being able to give up his collection of “vintage years of White-Out.” 😎

                    2. Oh dear, WP did an upgrade. Yikes!

                      I had an Olivetti portable in college that absolutely hated me. If I didn’t type with continuous, even keystrokes, it would throw in a space for the hell of it.

                      Since even, continuous typing was *not* one of my skills, I got the chance to do corrections a lot. With the special eraser and brush. Wite-Out(tm) wasn’t suitable for typewritten material until long after college, when that Olivetti was long in the rear-view mirror.

                    3. ’74 – ’79 was portable typewriter usage. I did have whiteout, used liberally. There is a reason why the 4 of us on our big graduation class project used a paid typist to prepare our formal write up. We weren’t the only ones. It was expected.

                      Also was the writer and editor of the school paper. How I ended up with it still confuses me. I did not volunteer. Every paper had a section that stated “Any complaints on sentence structure and spelling, just volunteered to be writer and editor of this paper.” Funny. No one did. Eventually someone did come along that felt they wanted the job. Gave them the mimeograph “paper” (whiteout was useless). Showed them how the mimeograph process worked and walked away.

                  1. “Blessed are the pessimists, for they have made backups.”

                    Yet another lesson: Make sure the automagic backup system doesn’t have any lingering beartraps. Several years ago, I set up resnapshot to back up my linux machine to a USB hard drive. But, I didn’t notice the option to tell it not to create the snapshot_root directory.

                    It worked fine until the backup drive went toes up. At shich point, the program thought I really wanted the backup to go on the main drive, and created the directory. On my small-as-possible root partition. Since the first backup is the biggest, the root partion was completely wiped out before the backup ended. Shit!

                    As memory serves, I had to go back to the distribution DVDs and reload the default packages. After which, I went to the config file and told it not to be so helpful. Then I got two USB drives, one for the day to day backups, and one that runs once a month just to be on the safe side.

                    It’s probably time to swap out the default drive; it’s been a few years.

                    1. A common mistake is to have a backup process defined to the last detail…. and a recovery process that has not been defined, let alone tested.

                      If you can’t recover it, why bother making it.

                    2. I’ve done a fair amount of recovery from those backups. In this case, it’s a matter of mounting the drive (kept unmounted unless doing a backup or manually mounted), going to the relevant backup directory, then going to the file. Then it’s a simple copy command for the file(s) of interest.

                      I also keep the shop computer as close to the house desktop as possible. 600′ away, so updates are via thumb drive and sneakernet. That one uses a SSD card for its backups; the shop is kept from freezing in the winter, but I’ve had DVD drives refuse to open in the cold. I’ll not trust a USB spinning drive out there. The main hard drive is OK so far (several years), but a SSD main drive would be an ideal replacement if necessary.

      2. “Principle #2 of Agile: Working software over comprehensive documentation”

        Which should read “Working software without documentation”

        8-(

        1. OTOH, I have read warnings that you should not debug comments, only code. At least you KNOW the code is doing what it’s doing.

          1. No. But comments can given an explanation on what the code was expected to do. Then when code doesn’t, it gives a place to for why. OTOH I’ve ran into code that did what the comments said it did but the client said it was “working wrong”. Depending on the situation, the client may be correct, or incorrectly informed.

            1. THAT’S usually a sign of “You gave me what I asked for but not what I wanted” syndrome.

              1. usually a sign of “You gave me what I asked for but not what I wanted” syndrome.
                ………………………..

                Not wrong. Except my last job where I constantly told end users “I am a programmer. Not an accountant. That is an accounting problem. This is how the program does it. Would you like me to talk to the boss? Or would you like to?”, pause listen. “Okay. Hold on a minute. I’ll go tell him what is going on, then transfer you.” Follow through. Or, more likely, because boss was on the road training and selling, “I will write this up, and pass it on to the boss.” The boss was an accountant. Sometimes the word came down to add a switch to accommodate the client. Usually not. More often than not, in this case, the end user was misunderstanding a state level requirement methodology. There were still a lot of requests for legitimate changes and enhancements (like the Obama health insurance year end reporting form to employees everyone gets these days. Not us, now that we’re retired, but son gets one. Heard the software is going to be printing pay checks. Shiver, I hate pre printed forms. Glad I’m out.)

              1. My last month. Client calls. XYZ department hates the change made. Put it back. Me: “Okay. But you paid for this change. It will be another charge to restore it. Sign the estimate and initial which way to do it. I’ll put it on the list.” Hint. Up charge for “do it right now”. “Let me know if you want the estimate.” Client “Let me get back to you.” (Never did. Before I left at any rate.) Guess which client was known for this? Constant problem when the clients insisted on having a go between for the end users. Not the only client who used IT go between. But they were the absolute worse for not having their clients, actual users, on the same page.

        2. The Reader first saw that at the Great Big Defense Contractor when one of our very senior software folks invoked it while explaining to a customer why Agile was a very bad idea for weapons software.

      3. The problem is that your code is obviously self-evident.

        The best time to comment code is about a month after you wrote it, when you can still bring it back in memory but it’s distant enough that you are able to tell what needs comments.

        1. I think it depends on the project load and complexity. The consulting gig I had needed to make code that was going to be supported by a team in Germany. One section of the code used an algorithm that was gloriously complex (not my algorithm, but I had to fit it in the program). It needed to be documented in detail for anybody to understand what in hell it was trying to accomplish.

          We were also developing the software in California and shipping working portions to the client in Bavaria. No way to put off documentation.

          As we were finishing the first program, the second one was ramping up. Since I was the lead codemonkey on both, I had to keep it commented before shipping, to maintain what sanity I had.

          1. No way to put off documentation.
            …………….

            Same. One person shop, ’90 – ’96. As one project was winding down the next was winding up. Either I documented then, or when something went wrong, or more often had to change, and I had to figure out what and why I’d done what I did, a year or two before.

            One section of the code used an algorithm that was gloriously complex (not my algorithm, but I had to fit it in the program). It needed to be documented in detail for anybody to understand what in hell it was trying to accomplish.
            ……………..

            ’96 – ’02. 100% B-Star buffered indexing, against a text data file (not my algorithm and as it turns out, not the original programmer’s either. Found code on internet and “made it work”. Never knew how it actually worked. Grrrrrr.) Ran into problems multiple times. First time just figured out how it worked, found problem, fixed it, after all won’t break again, right? Second time, reacquainted with how it worked, and documented the hell out of it (was not redoing the tracing, again). The last change I had to make the algorithm wasn’t broken, but the algorithm was the problem. Documentation paid off. Then documented that change and why I had to “break the algorithm” for specific condition. Bonus the comment details were there when I was called after being cut. Pointed that out. Caller wanted a “summary” of the comments (were extensive for a reason). No. They needed to read the comments.

    2. “I want a time machine so I can go back to when I wrote this undocumented, uncommented piece of crap and kick myself in the ass.”

      I was once tasked with ‘porting’ a ‘legacy’ program in 6802 assembly to a PIC30. There was maybe one short comment every 30 lines or so. Wasted a couple of hours trying to make any sense of it, then started on rewriting the program from scratch.

      (No, I did not write the original)

      1. There was a major program at one place that I worked that would had to be rewritten from scratch.

        There was an idiot COBOL command that “changed” Paragraph-Names that was used in that program so it was basically impossible to follow the logic flow of the program.

        Unfortunately, since it was a major program and apparently worked, nobody wanted to spend money and programmer-time to fix that garbage.

        Oh, that company shut down so I don’t know if it was ever fixed.

        1. Possibly the reason it was shut down had something to do with a “don’t fix it” attitude leaving tooth marks in their butts?

          1. There were plenty of other Upper Management Idiocies that are more likely to have destroyed that company. [Frown]

            1. I suspect we could all tell stories about those, and only some of them would involve software. 😦

      1. I know the feeling. I wanted to be able to work on something else.

        What I liked about the job from ’90 to ’96. My resume has a list of about 10 different programs and concepts, including the main multi-user system. When asked which part of each I did, some, like the main program were maintenance, and annual statistical growth model changes, as well as keeping the hardware running and upgrades. A couple were gut what wasn’t working, replace that code, add new concepts. But most were get concept from client, why, what currently doing, and create program, concept, design, code, deliver, train, and support. While the working tools, were generally the same, the requirements were not. Note, I was down playing what I’d done on the resume, for most programmers it was a huge list. I know for my managers they thought I should have been a automatic hire for one of the two companies diving up the divisions assets they bought. Obviously my managers were incorrect (they did try).

        Not so much on the subsequent jobs. But those jobs did not have the option to have as many different concepts to work on.

  3. I -so- need the Mal/airlock meme for at work. (Grin)

    …….

    Meme idea:

    Snarling Kzin warrior, facing the viewer in the attack, armed with a scaled-up AK-47 with bayonet. Caption:

    KalashiKzinti!

  4. Unfortunately, your defense lawyer rolled a 2 on her persuasion roll against the judge, so the evidence comes in anyway. A reroll saving throw will cost another six months to a year in prison, or you can take the plea deal and be eligible for parole immediately.

    I hate my job some days …

    1. Rolled under my “chutzpa” by 5.

      SchlemielMaus: “…. I never studied law….” schtick with perfect timing.

      (Boggles animator)

    1. I do wish we had our own Milei. Trump riles up the ruling class, but he’s not going to cut the government to anything like the extent Milei has managed.

      1. Before you blame Trump for that, you might want to look at the differences in basic law between them.

        1. I wasn’t blaming Trump exactly. My main problems with him there are that he, as a boomer, has a residual feeling that the system should work a certain way, which is hardly his fault, and that he was, at least the first time around, naive about what he was confronting. The latter, I think, has been largely burned away by the hysterical persecuation he’s been enduring.

          But if you laid out the basics of Austrian economics to him, I’m not sure how he would react, whereas Milei understands it in a way no other politician on the world stage does.

      2. Perhaps Mr Trump has been inspired to Bonsai the government.

        Small, highly constrained, and frequently pruned.

      1. I laughed so hard at kitty kitty bang bang. Then I tried to measure Jase for a belt. I don’t think I’ll need stitches.

      2. Back when Penguicon hadn’t gone full Jim Jones, I attended for several years. They had a ‘Geeks with Guns’ outing at a range. Those who went got badge ribbons which read “Bang! Bang!” It took far too long (ox slow…) to realize what I had to do. IThe next year I got there before the group left for the range and passed out ribbons which read “Chitty Chitty.”

      1. Well yes, but quite a few people will. What the left is really doing is trying to ban everything, so they can make exceptions through non-enforcement for “friends of the program” while using it as a cudgel to target opponents, often randomly, just to create an atmosphere of fear and distrust. That is pretty much the leftist totalitarian game plan as practiced by every leftist dictatorship in history.

      2. via PJ Media:

        https://pjmedia.com/marktapscott/2024/03/02/another-embarrassing-mainstream-media-christian-nationalism-moment-n4926952

        A reminder that the left rejects even the Declaration of Independence because they believe rights are granted to people by The State as a privilege, rather than are natural and inherent in each person. If rights are inherent, either God given or for those who don’t believe in God, just plain natural, then The State wrongly takes away and violates the rights of people’s when it deprives them of those rights. THAT is what the left finds intolerable. For them it is the “collective” uber alles, with the “collective: needing to be absolutely ruled by a small group who “know what’s best” for everyone. This was the essence of Obama’s repeated statements about knowing what various people wanted better than they themselves did.

        1. This was intended as a new post, but apparently once someone replies to a comment, it is adding those new posts as replies to the first post one made. WP Delende Est.

    1. A standard joke was that a group would be so “conservative” or reactionary that they disapproved of fire. Which makes the “Greens” ….

    2. When they banned lighters and matches on airliners, I provided Pop with a tinderbox with “burning glass”, flint chips and steel, tinder, and a pipelighter (slow match rope with brass band holder/extinguisher.

      Not prohibited, oddly enough.

    3. True story: I was at summer camp a couple of years back and the Wilderness Survival folks came back from their overnight verifying that a scout had managed to successfully start a fire by banging two rocks together. Not flint; this was entirely granite. Not by light or friction of wood on wood. Two. rocks.

      Good luck with that. We teach scouts how to start fires in ALL manner of ways. It is literally one of the oldest human skills. It’s a PRE-human skill.

      1. teach scouts how to start fires in ALL manner of ways.
        ……………

        Yes.
        One match fires.
        Flint and steel.
        Rocks.
        Two sticks, with and without bow string.
        Steel wool and batteries.

        Chemicals, or at least demonstrations, “it’s magic”. Order of Arrow scouts even set up the latter fires to go. (When one of the adult scouters are chemists. The starting fires at district and council events were colorful.)

        I’m missing a few methods. I am sure.

        1. I learned most of those from watching survival shows on cable. The shows may have a lot of staging, but they did and do teach some useful things from time to time.

          1. It is rather important to practice firemaking methods before the life-threatening event that demands an improvised fire.

            1. Practice. Practice. Practice. Then Practice some more. Just seeing fire making being done and memorizing the technique is not not enough to actually get a fire going. As many a scout and scouter (adult know it all) have learned. Getting a fire started and burning does not guarantee an edible meal should one have to cook over a wood fire. One of the reasons why first camp out with new scouts we encouraged parents to participate. We have parents and scouts separated. Showing each what the scouts are going to be working on their first year. Then we have the parents try to build a fire, and cook an easy meal. Because being adults they should be able to do whatever their first time, right? (Shh B. Durbin.) Unless parent is former scout (older GSA or BSA) themselves, or former military. Results are hilarious. Point is to be able to state, and hopefully get it to sink in, their scouts need to practice. For most that it is during outings. Even that is getting difficult because even locally that rules out summer outings. OTOH if they can build a fire in pouring down rain, they’ve learned to build a fire.

      2. Sadly I think the Scouts are another institution that that the left has gutted and skinned and wears like a suit parading around looking like the original but utterly antithetical to it.

        1. YMMV

          Disagree. From the outside might look like it. But, although currently on the outside, know enough of the inner workings to know, not so, especially at the unit, district, and (at least our smaller) council, levels. Do not ever remember checking sexual orientation on any scouting forms or applications (started as adult scouter ’95, as six year old entered cubs, last year registered 2010). Biological sex on forms, yes. The idea of anyone, whether spouse was also a scouter, discussing personal bedroom activities, would have been inappropriate and gross. I don’t care what type of partner one had.

          The allowing females to join boy scouting? As far as I’m concerned at least 60 years too late.

          Granted outside looking in, but Girl Scouts and Campfire are what “the the left has gutted and skinned and wears like a suit parading around looking like the original but utterly antithetical to it” a long time ago. So, pretty sure that is inaccurate too.

          The Eagle (BSA), Gold Award (BSA Venture), Gold Award (Girl Scouts), Wehelo (Camp Fire USA), are all equivalent in that they all show leadership, but they are not the same, in methods. BSA Eagle and Venture Gold achievements are closest. But then both are under the BSA umbrella. Age differences: BSA – 10 1/2 – 18th birthday, Venture – 14 – 21st birthday.

  5. All good (as usual), but…

    “Not An Old Bus” – AARGGGHHHHH!

    “When it’s carping time, at the old home place…”

  6. In video game news…

    If you’ve been paying attention to video games over the last few months, you might have heard of an outfit called “Sweet Baby Inc.” They work as a “consulting” company for video game developers that want to improve the “diversity” in their video games. While some had heard of them, most were unaware that they were affiliated with such recent releases as Spider-Man 2, Horizon, and the Saints Row reboot.

    And then a AAA title called “Suicide Squad Kills the Justice League” was released about a month ago. The game flopped massively. It was such a huge flop that it might take down the development studio that released it. And the game has SBI’s fingerprints all over it. So, suddenly, SBI is in the spotlight.

    The PC video game platform and market Steam has what are called curators. These are individuals who post short reviews (usually a couple of sentences) of video games that they may or may not have played, along with a quick “Recommended”, or “Not Recommended”. Sometimes the reviews are serious. Sometimes they aren’t (there’s one curator called “Hodor” that just has the word “Hodor” for each of his – or her – reviews). In the aftermath of the Suicide Squad debacle, a new curator has appeared. He calls himself, “Sweet Baby Inc. Detected”.

    Sweet Baby Inc. Detected does exactly what his name states. He identifies games that Sweet Baby Inc. was involved in, and then lists “Not Recommended” next to each of them. This allows players to recognize games that Sweet Baby Inc. was involved with, and avoid them. That’s it.

    The response from Sweet Baby Inc. has been predictable. Employees have accused the curator of “harassment”, and attempted to launch campaigns to get the curator blacklisted and banned from Steam for supposed (and non-existent) violations of the Steam code of conduct.

        1. “Don’t be a Jerk.”

          “But he was a jerk first (toward me).”

          IE Humans always find excuses to do Bad Things.

      1. Codes of conduct serve good purposes. They’re generally “Have a modicum of courtesy, or else,” rules. They exist to establish bounds that people understand they can work within. This is useful when you have people coming together from a wide array of backgrounds.

        The problem is that certain people will try and twist anything and everything. And the “Ends justifies the means” attitude that many have these days means that such people will use the most insane pretzel logic to rationalize why their target violated the rules.

        After all, we live in a world in which people will claim with a straight face that the act of breathing by a white person is an act of white supremacy.

      2. I had a reply to this, but it disappeared…

        Codes of Conduct are needed when you have a large number of people coming together from a wide range of backgrounds. Different people will disagree over what is and is not appropriate behavior. Codes of Conduct lay it all out.

        The problem isn’t the codes. The problem is people who try to twist logic into a pretzel in an attempt to justify what they’re doing. Individuals who believe that the ends justify the means (including many of the Woke types) usually fit this description.

        1. The problem is that the code of conduct is only ever enforced by the Left, for completely political reasons, often in direct violation of the actual words of the code of conduct.

          1. No, the problem is WHEN that happens.

            I have seen places that actually use the Code of Conduct to set expectations, and apply them.

            Of course, the biggest place that had rules, and actually applied them, I actually paid for. It’s called Ricochet.

            That is in the past tense because they decided that it was more important to encourage a wider range of voices, and folks on the progressive side were literally incapable of staying on the site if the rules were enforced evenly.
            This was made very obvious and open when a paying customer was suspended from the site for quoting someone who had attacked him. The attacker was not even scolded, much less punished.

            1. This. For example, X is more even-handed now than Twitter was, with Musk going so far as to declare that using “cis-” as an epithet can get the user into trouble. Given that only progs use it as an epithet, it’s evidence of a shift toward even-handed policing of the Code of Conduct.

        2. I was debating responding as well.

          The “don’t be a jerk” thing has been weaponized as much as Code of Conduct– because “jerk” means “I feel bad, you must be punished.”

    1. Look, this is very simple:

      Karens have the divine right to tell everybody to stop having fun the wrong way, which is any way that is actually fun.

      And everybody else is Not Allowed to do anything that interferes with that divine right.

      1. Please see the ‘Don’t Brake For Communists’ Meme, and apply accordingly. snicker

      2. Weirdly enough, the use of the name Karen as an epithet has had its worst effects on actual sweet people named Karen. (There are also kind people named Karen who are self-assured enough to not be hurt by this; we’re talking about the folk who are getting harassed regularly because their parents weren’t prophets.)

        The solution is simple. The word “harpy” exists and has a roughly congruent meaning. I propose that instead of calling someone a Karen, we revert to the term “shrieking harpy” and call it good.

      3. Karen doesn’t believe in divinity, therefore her authority must stem from another source.

      1. Link doesn’t work.

        Given that the old people are less likely to be invested in this sort of nonsense, not hiring them makes a certain amount of sense…

      2. Something got cut off – the error is:

        MissingKeyMissing Key-Pair-Id query parameter or cookie value

  7. Way back when the Y2K scare was ramping up I had to “baby sit” an old retired programmer who was contracted to “fix” our state jail system and it’s data. The original had been done in Assembly language and nobody could read it. We did find a tiny bit of paper documentation in an old rusted file cabinet at the archive facility but it didn’t help much. He cleaned it up, it (along with all of us) made it through Y2K and last I knew (about 2012) it was still running. All the (massive/expensive) efforts to get something else to actually work failed from 2000 on – after that, I left and just don’t care. Yeah… we were still using old main frame ‘dumb terminals’ for it too.

    1. I was specifically hired in 1990 to deal with up coming Y2K. Fix had to be completed before end of 1990. COBOL program with no documentation. So much fun to untangle.

      1. Amusingly enough, yesterday my husband was drooling over the phone he wants when his current one dies.

        It’s a flip-phone.

        The one with the folding screen, but he wants it because it’s a fliphone, too.

        1. Me too. Samsung Folding Smartphones. I want this version (whether I can get beyond the “ouch, ouch, ouch, ouch” price point, is yet TBD.

          An aside. “Oh. Look. WP has changed how things work. Again.” OTOH pictures show up before you post.

          1. That’s the very one we were drooling over!

            Hm….
            :pokes, sees if can upload from PC; cannot. Sees if can upload from Reddit:

            1. sees if can upload from PC; cannot
              …………………………

              Well. Dang. That is too bad.

  8. Re the elephant meme: I bet that’s Indian elephants. African elephants look at people and think they’d look better flat.

  9. Hi, Sarah. I’m a daily visitor and have purchased numbers of your books, especially the Dark Ship and Shifter series. Quick question from the author of exactly one book: why am I finding my book for sale on Barnes & Noble, Kobo, Everand, and Apple, when I published it in 2015 on Smashwords.com and then ignored it completely? It’s a general history of the age of power, starting with Hero’s invention of the impulse steam turbine, two millennia ago in Alexandria. I never expected it to sell to anyone but fellow gearheads, and can’t understand why it is being offered in these new venues without my knowledge. Thanks!
    Bruce Abbott

    1. I haven’t been back to Smashwords for a ehile, but when I signed up they were a dustrubution channel, soreading your book out to lots of other sites.

      But not to Amazon.

      1. Found it on Smashwords. “Since Hero: A Mechanic’s History of the Age of Power” Amazon gives no results.

  10. This seems like the closest thing we get to an open thread, and someone posted this on instapundit and I wanted to share it here.

    THE WRATH OF THE AWAKENED SAXON
    by Rudyard Kipling

    It was not part of their blood,
    It came to them very late,
    With long arrears to make good,
    When the Saxon began to hate.

    They were not easily moved,
    They were icy — willing to wait
    Till every count should be proved,
    Ere the Saxon began to hate.

    Their voices were even and low.
    Their eyes were level and straight.
    There was neither sign nor show
    When the Saxon began to hate.

    It was not preached to the crowd.
    It was not taught by the state.
    No man spoke it aloud
    When the Saxon began to hate.

    It was not suddently bred.
    It will not swiftly abate.
    Through the chilled years ahead,
    When Time shall count from the date
    That the Saxon began to hate.

  11. Occasionally, the Lord sends us cause to howl like a hyena. This is Houston’s Democrat DA.

    https://www.bizpacreview.com/2024/03/06/clerical-error-harris-county-da-told-her-ballot-had-already-been-cast-when-she-tried-to-vote-in-her-race-1442969/

    “A Texas district attorney who was unseated in the Democrat primary on Tuesday reportedly had issues when casting her own ballot.

    Harris County District Attorney Kim Ogg conceded to Sean Teare, a former prosecutor in her office, on Tuesday night. But earlier in the day, Ogg was “raising alarms about voting procedures there after she was turned away from the polls early Tuesday morning because her ballot had already been cast,” CNN reported.”

Comments are closed.