Monday, August 31, 2015
Sometimes I say that someone is re-inventing the wheel.
But, from me, that can't be an insult.
You see, if I could find someone to sponsor me, I'd be re-inventing the entire information industry infrastructure.
Information encoding? Yeah. I want to replace Unicode with an encoding standard that would encompass the functionality of asn.1 and sgml in one rational, accessible whole, not to mention make embedding binary data in text work much more smoothly. And reduce or eliminate glyph aliasing with out-of-context characters. And separate the international encoded sets from the national encoded sets, to help reduce such aliasing and make regular expressions work better in local contexts.
That's definitely re-inventing a lot of things perceived as wheels, not presently worth the attention of further refinement.
Programming languages? Yeah. I want to re-invent the language C, the runtime, add a couple of storage classes to reduce the problems of overwriting local variables and controlling concurrent access from separate threads, and add little bits to function declaration and call syntax. And I want to reinvent a language called FORTH, so it would be more amenable to being used as a user interface shell language, among other things. And re-invent Unix with a new executable object format supporting all this. That's going to be equivalent to an earthquake in userland, not to mention in the system itself.
Networking? Of course. I want to get rid of IPv6 and implement nested IPv4 addressing. Make NATted addresses optionally visible externally, to open up more static addresses and reduce the incentive for ISPs to charge through the nose for a static address, for starters.
CPUs? Those too. Intel has been burning up resources building their monopoly on the CPU market for far too long. ARM helps, but too many manufactures are too willing to play games trying to lock their customers in. And no one really supports proper separation of user resources in current CPUs. We need to focus away from raw speed and more on stability and securability.
And so on.
Yeah, I want to re-invent wheels. So, if I merely note that someone is re-inventing wheels, that, in and of itself, is not evil. And I do not intend insult by it.
Re-inventing wheels is good for many reasons, and not just to provide churn for the sales crew to work.
(Re-inventing wheels solely for sales churn is somewhat evil, but it can be better than keeping the world as it is. I should rant about that sometime, too.)
Friday, July 17, 2015
Thursday, July 16, 2015
I was at Kinokuniya on a Friday night a couple of months ago looking for some new teaching materials and thinking things like, "I could make something better than this if I had the time." and "Would this save me enough time to be the limits it imposes?"
And I found a novel from deep in my past.
A really good friend gave me a copy of the original novelette more than twenty-five years ago, so I picked it up to read a few pages. I ended up reading it until closing time, buying a copy, reading it on the train, and finishing it, standing up in the light by the ticket machines at my station, before I went home at about 3:30 Saturday morning.
[One of the less important thoughts I had was to compare the relationship of the original novelette to the novel by offering a parallel with Arthur C. Clarke's Against the Fall of Night and The City and the Stars. But all that really does is to point out that authors do, sometimes, re-write their own works.]
I get lonesome for English, and, in truth, if I don't read, my English goes downhill.
But I can put Harry Potter or Heinlein or even Bradbury down. [Or Clarke.]
Ender's Game is a good book, very thought-provoking, and not badly written at all.
Orson Scott Card's prose is reasonably clean, so it should also be a good novel for non-native students of English, as well.
[I have noticed a trend, as personal computers allow people to write more and write more quickly, away from the careful editing and honing that used to be a part of the publishing process.
Scott acknowledges the imperfections in the novel. Some of them were partially addressed, to varying effect, in producing the movie.
No literary work is perfect. So what? Go read Wuthering Heights, and think about what we'd miss if we insisted on perfection.
I'd still prefer to use Word Perfect 3 for serious writing, rather than any of the word processing software we have available today, but that thought has nothing to do with Ender's Game.]
The novelette and the novel are not the same. Some differences are minor, some not so minor. But it's the same story.
I also rented the DVD. Showed it to my daughter. She thought it was interesting, but not as interesting as her anime. She isn't going to suddenly decide to try to read the novel in English, at least, not until she's finished with The Wizard of Oz.
The movie is significantly different from the novel. Some of the differences are substantial.
[But they do not interfere one of the more important messages of the book, one which becomes more and more important as our modern society becomes ever more "modern". Bad things happen when people manipulate people, even with good intent. And then we have to let the world go on without what might have been.]
The DVD shows some cut scenes that I think I agree should not have been cut. It might have made the ending more readable, but the ending should be readable to any serious science fiction fan.
[Those scenes would have made the movie more satisfying, and more understandable, I think. Spoiling the ending should not have been a concern.]
Movies do not always have to have O'Henry-style surprises.
[I think I want to emphasize one message from the book: Failing banks, failing nations, marketing wars that destroy the foundations of the industries the markets were founded on. Why are we so sure have to win it all?
And I recommend the sequel, Speaker for the Dead, as well. It is not the usual more-of-the-same that you have in sequels.]
Thursday, July 9, 2015
When the site lists its fingerprint as something like
SSH2/RSA 2048 86:7b:1b:12:85:35:8a:b7:98:b6:d2:97:5e:96:58:1d
(Note, I am not trying to mirror keys here. If anyone reading this needs the keys for some site, such as the one I linked to above at one point, that person should go to the site itself, and complain loudly if keys can't be found.)
That format is the old, less secure MD5 format.
Go into the ssh configuration file for the user, probably something like
and add or uncomment this line:
But be sure to comment it back out when done, so you use the more secure protocol options instead.
(Should edit this when I'm awake again to add the site specifier line, which partially mitigates the problem of choosing the less secure protocol options. And otherwise say more sensible things.)
(Also need to complain loudly to said site about not publishing the SHA256 fingerprint keys yet.)
Wednesday, July 8, 2015
I just posted this morning about big companies behaving like Microsoft, fumbling the technologies and blaming on the competition.
I have a USB flash drive which I use to transfer files at work. Over the last few days, I have used it to transfer files between a MSWindows7 notebook machine on my desk and an MSWindows7 narrow tower machine up in the computer lab. I always get nervous about moving data on USB drives like this. Microsoft's software has bitten me in the past. So I back it up periodically.
But it had been a while since the last time. Have I backed it up this week?
So it was a little bit of an unpleasant surprise to plug it in to the notebook on my desk and see the warning dialog:
"Some problems have been detected in several files on this media."
"Do you want to fix them?"
There is no right answer to that dialog box.
Hit "Yes" and it messes up your file system.
Hit "No" and it messes up your file system a different way.
Close the accursed dialog without answering and it messes with your file system for spite.
Rip the USB out without mounting, or try to leave the dialog unanswered while you try to properly unmount the USB, and it leaves the file system and/or partition structures in a half-altered state.
It shouldn't be writing anything to those structures at that point in the first place. You don't have to write to those structures if you are only reading files on the disk.
Yes, Microsoft screwed up their various definitions of FAT file systems and the associated partitioning standards.
And they patented their screw-ups and charge people who produce non-MSWindows OSses patent-use fees for not using their patented standards screw-ups.
And if you refuse to pay because you don't want to use their intellect-impaired "Intellectual Property" anyway, they take you to court and force you to provide all sorts of proof that you and you grandparents and your in-laws have never used those patents, are not using them, and never-ever-ever will, as long as everyone lives.
Until you just give up, sign on the dotted line, and commit yourself to refrain from doing anything that would ever harm Microsoft's position in the market-place, etc.
(No, I'm not exaggerating, and I'm only speaking metaphorically about the in-laws and grandparents. Think technology relationships. That is what Microsoft's "right to innovate" is all about.)
But that doesn't mean they have any reason to write to your partition map just to mount and read the media.
So, I have to capture the data on this drive when I go home tonight, and reconcile it with my backups. Then re-format and figure out what I need to carry with me. The fewer files I carry, the safer it is.
And, incidentally, a pox on Intel, too, for helping bad players in the industry impose these poorly designed, fragile USB media on us. A pox on Microsoft and Intel both.
About a month back, my aging primary computer died. I think the CPU needs more heat-sink grease, but it's a 32-bit Sempron. The RAM is kind of tight, too, just 760 MB. Using the software everyone around me thinks I have to use has been getting difficult. So I am moving my operations to other hardware.
My mail client, when I'm not being lazy and logging in to my gmail account from the web, is usually Sylpheed. (A little more about Sylpheed from wikipedia.) I had been using Sylpheed on a regular basis when things hit the fan with systemd, but I had been rather lazy from that point until the hardware died. Moving from debian and learning how to use openbsd didn't leave me much time for setting up the mail client. I don't know when I last logged in to gmail with Sylpheed.
After moving to new hardware a few weeks back, I find myself unable to log into my regular gmail account from sylpheed. (I can log into my family gmail-hosted e-mail, however.)
Couldn't find any answers on Google. Didn't have time to dig into it.
I've filtered notices and other junk from Google into a folder just for Google, and I often forget to look there. I have to start looking there more often.
I looked in there just now and found a notice from Google from a few weeks back:
We recently blocked a sign-in attempt to your Google Account [address].
Sign in attempt detailsDate & Time: [date and time]If this wasn't you
Location: [location]Please review your Account Activity page at https://security.google.com/settings/security/activity to see if anything looks suspicious. Whoever tried to sign in to your account knows your password; we recommend that you change it right away.If this was youYou can switch to an app made by Google such as Gmail to access your account (recommended) or change your settings at https://www.google.com/settings/security/lesssecureapps so that your account is no longer protected by modern security standards.To learn more, see https://support.google.com/accounts/answer/6010255.
The Google Accounts team
This email can't receive replies. For more information, visit the Google Accounts Help Center.
You received this mandatory email service announcement to update you about important changes to your Google product or account.
So I was, indeed, blocked.
Well, if I were using MSOutlook or MSOutlookExpress, especially on MSWindows98 or MSWindowsXP, I'd almost consider this a reasonable approach to a real problem.
Except I can log in to gmail from my Android 4.something-now-old tablet that is not getting updates from anyone any more. That fact puts a little unsavory perspective on things.
I checked the links. It was more of the same. Apparently no comprehension that it might be Google fumbling the more secure login handshake.
This reeks of Microsoft's deliberate habit of fumbling the newer, better open standards in an effort to convince everyone to believe they have to use Microsoft's less secure, less capable, completely inferior offerings.
(Financially speaking, Microsoft has been quite successful at this game. Four of five people who read this post will still be running some sort of Microsoft OS, and are still in the habit of thinking that Microsoft Office is the definition of office productivity software. What OS are you using? Maybe you are using Firefox, but who makes the rest of your software?)
I probably need to be explicit about this, but Sylpheed is not a fly-by-night Android App built by someone I've never heard of. The OS is not that Android OS, that, finally, in version 5, supports some semblance of one of the foundations of a secure OS, a way to login as a specific user.
(... But is ultimately still effectively tied to the OEM's opinion of what freedoms the end-user should be able to exercise -- Breaking out of the OEM's jail requires, for most people, using another fliy-by-night app by some developer you-nor-I have ever heard of, only mechanically rated by the PlayStore rating system.)
So much for Google's dalliance with "Don't be evil."
(And, yes, this one does get posted, even though I'm much calmer now than when I started typing. Filtering a few swear words and other less-coherent rants out of the post helps me calm down. I have a lot of these rage posts that I don't bother posting. There are more that I delete. But this one gets posted.)
Friday, May 29, 2015
I installed clamav on my openbsd notebook because I decided I needed clamav for when friends ask for help.
Trying to get the current database resulted in the message
ERROR: Please edit the example config file /etc/clamav/freshclam.conf
ERROR: Can't open/parse the config file /etc/clamav/freshclam.conf
A quick scan through the configuration file didn't reveal any particular problems that I noticed. (If it had been a snake, it'd have bitten me.) No file ownership or permission issues.
Searching the web sent me to lots of answers like this:
Comment out the "Example" directive on the third or fourth line of the file.
The clamav team really wants you to at least look in the file, I guess.
Not a bad idea.
I'm downloading the current database now. Next step is to write an image of this to the bootable USB drive.