The Old C Dog

Programmer. Techie. Problem solver. And you would be too if you were me.

Also, if I'm the old C dog, why is my profile picture a cat?

The Old C Dog boosted:
Herbert "Bracket" Wolversonherberticus@fosstodon.org
2024-05-23

My new book – Advanced Hands-on Rust – is now in beta! Intermediate to advanced Rust, writing games in the Bevy engine. Generics, traits, macros, library creation and more. pragprog.com/titles/hwmrust/ad
#rust #rustgamedev #bevyengine

@pragprog beta books get you most of the e-book now (sadly, we can’t ship out a new print for every beta!), and each subsequent release as the book advances towards final release. You can also file bug reports and help make the book better.

The Old C Dog boosted:
2024-05-20

Microsoft Open-Sources the original source code for GW-BASIC from 1983 devblogs.microsoft.com/command

2024-05-17

I saw a post that started, "The most dangerous thing about ChatGPT is..." and finished by describing a situation where ChatGPT gave a confident seeming and incorrect answer.

To me, the output is indistinguishable from the glib, confident, yet incorrect statements made by politicians and people who are wrong on the internet.

So, let's finish that sentence. IMO, "The most dangerous thing about ChatGPT is that people actually listen to it."

No. I won't debate you on this point.

The Old C Dog boosted:
Fi 🏳️‍⚧️munin@infosec.exchange
2024-05-12

The tendency for corps to automate away low-tier work has some unfortunate consequences when this tendency hits things like SOCs - at that point, removing e.g. tier-1 triage type work produces minimal cost savings (junior analysts aren't all that expensive) but also removes the work experience that is required to learn how to become a senior analyst.

When your educational pipeline is interrupted, this also interrupts knowledge transfer from seniors within the organization, resulting in increased institutional knowledgebase decay over time.

Many of the services currently in production are highly resilient to disruption - and well they should be; they've been built from the accumulated knowledge and expertise of many people over a period of years, with the explicit goal of creating resiliency, maintainability, and performance.

However, that resilience is a finite resource for these services, and is maintained over time by the application of institutional knowledge and expertise; even the best-documented and most-resilient systems will decay over time if the personnel who comprehend its structure and maintenance requirements depart the service-operations role without passing this distillation of knowledge down to their successors - successors who, in turn, need an onramp to obtain such roles in the first place, an onramp that was formerly expected to be the tier-1 roles that are being progressively automated away.

And that's not to say that all tier-1 roles will vanish, and certainly not immediately - but the progressive reduction in their availability will impact the pool of personnel available for these roles.

Like all complex systems, the ones I'm talking about aren't going to immediately fall over; instead, we see gradual losses in effectiveness over time - for "the system by which security analysts are trained from neophytes to senior level" this will look like a much-reduced qualified candidate pool over time, with fewer and less-qualified candidates for the jobs that are available.

This is likely to end up being compensated for in the short run by outsourcing and exploiting international workers from areas that are lagging behind in technological availability of these junior-removing automations - but these workers will necessarily not have the same context and institutional knowledge that in-house workers will have about the nature and purpose of the systems being used, which will cause impedance mismatches that we already see with, e.g., MSSPs and their clients.

I don't have a clear feasible solution for this. I'd like to see various outfits that are working to automate away junior positions engage with this directly, tho, and to have frank and open discussions about ways to teach people how to work with their systems to gain that expertise so that they can be effective seniors in the paradigm that those companies envision in the future - tech requires people to operate, and if your product is intended -for- the disruption of such a market, describing the new system that you envision resulting from the disruption is, I think, a requisite for ethical participation in such markets.

Ultimately, I think that considering the whole lifecycle of how -your- product interacts both with the system as it is -and- with the system that your product will define with its success is absolutely necessary for a product to remain successful over time.

The Old C Dog boosted:
mhoyemhoye
2024-05-11

I'm on here looking for text indexers and everything is 'lightning fast exoscale terafloops that scales to enterprise quantawarbles with polytopplic performanations' and it would be great if this industry could breathe into a bag until it remembers that one person with one computer is a constituency that matters.

2024-05-05

The fourth installment of Let's Build a Virtual Machine. In this post, we add much-needed labels to the assembler.

The Old C Dog boosted:
2024-05-04

A busy week, but I've completed all of the major components of the Virtual Hagelin M-209! The print head mechanism is working & cage indicator modeled along with several other levers. Still need to mount & animate then lastly, work out how to make it print!
#tnmoc #bletchleypark

2024-04-28

@swan_tower Heaven's Vault is Chants of Sennaar with a better story.

2024-04-15

@fluidlogic @virtualcolossus got to love computers that did pounds, shillings and pence. I wonder if this is from the same family as the EMIDEC? emidec.org.uk/emih0hom.htm

2024-04-14

"Blazingly fast" should not need to be said. It should be the default, along with "secure" and "simple".

This is not knocking any product or project that proclaims these qualities.

It's the sad realization that "fast", "secure" and "simple" are qualities that users of our products should be able to take for granted, but can't.

We, the software industry, can do better.

The Old C Dog boosted:
2024-04-13

Any experienced programmer worth their salt will tell you that •producing• code — learning syntax, finding examples, combining them, adding behaviors, adding complexity — is the •easy• part of programming.

The hard part: “How can it break? How will it surprise us? How will it change? Does it •really• accomplish our goal? What •is• our goal? Are we all even imagining the same goal? Do we understand each other? Will the next person to work on this understand it? Should we even build this?”

The Old C Dog boosted:
mhoyemhoye
2024-04-11

Talking to @grimalkina about mathematics in computing, I'm reminded of the fact that all the theory courses that teach you all the different ways you can think about computer programming happen _after_ all the weeding courses purpose-built to cull out anyone who, y'know, actually thinks about programming differently, and just wondering how many human beings this field will casually and structurally throw overboard so it can stay a narrow, incurious priesthood.

2024-04-11

If I see "Type /ai for writing assistance" one more time then I will be sending a strongly worded letter to the Atlassian CEO.

2024-04-08

@SteveSyfuhs @tiraniddo I'm very much looking forward to getting into this in a company learning day tomorrow.

2024-04-07

I still can't quite believe an instruction set that handles pounds, shillings and pence.

emidec.org.uk/emihinst.htm

The Old C Dog boosted:
Greg Wilsongvwilson
2024-04-07
2024-04-07

The third installment of Let's Build a Virtual Machine

The Old C Dog boosted:
2024-04-03

Please tech companies, keep adding AI to your services as a "premium" paid for feature that allows me to avoid it and also not have to pay you. Win-Win.

2024-04-02

The second installment of Let's Build a Virtual Machine

The Old C Dog boosted:
Liam O'Mara IV, PhDLiamOMaraIV
2024-04-01

My father sent this to me, and I figured it belongs here with you fine folks. 😀

Client Info

Server: https://mastodon.social
Version: 2025.04
Repository: https://github.com/cyevgeniy/lmst