General > General Technical Chat

An observation on homework problems

<< < (9/23) > >>

paulca:

--- Quote from: coppice on August 13, 2020, 01:12:32 pm ---
--- Quote from: tggzzz on August 13, 2020, 12:52:29 pm ---You couldn't be more wrong, about me and about JSP.

I investigated JSP (and the other contemporary techniques) in the mid 80s, and rejected it as boring and not the way forward. My decisions and choices at that time have been proven correct.

--- End quote ---
JSP was just another cult. People came back from the JSP courses like religious converts, but later realised they'd got very little from it. The mid 80s was a little late to be looking at Jackson, Yourdon, etc. They had already passed peak BS by then.

--- End quote ---

Em.  All CPUs still function purely on JSP. 
Sequence, selection, iteration.

At least that's what I took from my exposure to JSP.

Even a concept as ingrained as a C function has no real model in a CPU.  The stack and branch operations required to deliver it in assembler are fabricated into sequence, selection and iteration operations.

That is what I meant by it not mattering what paradigm or language you use, it all ends up at basic JSP anyway.

Granted there is the rest of JSP stuff.  It's been decades since I looked into it.  Was it not just top down design?  Modularise, divide and sub divide into smaller chunks to work with?

paulca:

--- Quote from: tggzzz on August 13, 2020, 12:52:29 pm ---And still people don't understand the "fallicies of distributed computing". They think because the framework documentation omits to mention them, the problems have been solved. Ha.

As for people understanding the limits defined by Lamport, or the Byzantine General's problems, or the split brain problem, they have no clue. They think FSMs are something to do with compilers.

--- End quote ---

I suppose it depends on what you want to achieve by using distributed computing.

If you are simply in the "compute" space aiming for speed/throughput/latency it would be completely different to the scalability requirements and presenting dynamic platforms to allow basically architecture as a service.  aka Kubernetes.

Similarly, the Big Data concepts in distributed computing have allowed data mining and processing on a different level which simply was not achievable until recently.

Maybe you mean the age old confusion that management might have that if 1 machine takes an hour to do a job that 2 machines will take half an hour.  That's just management being fucking dumb.  I don't think I have met a manager with a degree in anything related to management, so I normally treat them with caution ... tolerate them or use them to my advantage.

tggzzz:

--- Quote from: paulca on August 13, 2020, 01:35:47 pm ---
--- Quote from: coppice on August 13, 2020, 01:12:32 pm ---
--- Quote from: tggzzz on August 13, 2020, 12:52:29 pm ---You couldn't be more wrong, about me and about JSP.

I investigated JSP (and the other contemporary techniques) in the mid 80s, and rejected it as boring and not the way forward. My decisions and choices at that time have been proven correct.

--- End quote ---
JSP was just another cult. People came back from the JSP courses like religious converts, but later realised they'd got very little from it. The mid 80s was a little late to be looking at Jackson, Yourdon, etc. They had already passed peak BS by then.

--- End quote ---

Em.  All CPUs still function purely on JSP. 
Sequence, selection, iteration.

At least that's what I took from my exposure to JSP.

Even a concept as ingrained as a C function has no real model in a CPU.  The stack and branch operations required to deliver it in assembler are fabricated into sequence, selection and iteration operations.

That is what I meant by it not mattering what paradigm or language you use, it all ends up at basic JSP anyway.

Granted there is the rest of JSP stuff.  It's been decades since I looked into it.  Was it not just top down design?  Modularise, divide and sub divide into smaller chunks to work with?

--- End quote ---

Looking at https://en.wikipedia.org/wiki/Jackson_structured_programming#JSP_and_object-oriented_design it does look boring and trivial, and it looks like you are right.

It seems as if I may have been thinking of JSD https://en.wikipedia.org/wiki/Jackson_system_development

It has been a long time, and I've no regrets purging them from my memory :)

tggzzz:

--- Quote from: paulca on August 13, 2020, 01:46:28 pm ---
--- Quote from: tggzzz on August 13, 2020, 12:52:29 pm ---And still people don't understand the "fallicies of distributed computing". They think because the framework documentation omits to mention them, the problems have been solved. Ha.

As for people understanding the limits defined by Lamport, or the Byzantine General's problems, or the split brain problem, they have no clue. They think FSMs are something to do with compilers.

--- End quote ---

I suppose it depends on what you want to achieve by using distributed computing.

If you are simply in the "compute" space aiming for speed/throughput/latency it would be completely different to the scalability requirements and presenting dynamic platforms to allow basically architecture as a service.  aka Kubernetes.

--- End quote ---

It always depends :)

I've never bothered to look at kubernetes, so I can't comment on that.

However, one system I architected in 98/99 presumed globally distributed order fulfillment with rapidly evolving products and fulfillment mechanisms.

Others were telecom systems, where latency, horizontal scalability and availability are key. And the telco system changes without notice - even the network operators know they don't know what's in their system, let alone other companies' systems :)

Yes, that did involve evaluating and implementing both clustered frameworks and platforms (J2EE, Spring, JAIN SLEE), and custom coded servers.


--- Quote ---Similarly, the Big Data concepts in distributed computing have allowed data mining and processing on a different level which simply was not achievable until recently.

--- End quote ---

Indeed, but there have been very few fundamental advances there.


--- Quote ---Maybe you mean the age old confusion that management might have that if 1 machine takes an hour to do a job that 2 machines will take half an hour.  That's just management being fucking dumb.  I don't think I have met a manager with a degree in anything related to management, so I normally treat them with caution ... tolerate them or use them to my advantage.

--- End quote ---

I've seen all sorts of dumbness, technical and managerial. I'm glad I'm out of all that!

coppice:

--- Quote from: tggzzz on August 13, 2020, 02:20:34 pm ---Looking at https://en.wikipedia.org/wiki/Jackson_structured_programming#JSP_and_object-oriented_design it does look boring and trivial, and it looks like you are right.

It seems as if I may have been thinking of JSD https://en.wikipedia.org/wiki/Jackson_system_development

It has been a long time, and I've no regrets purging them from my memory :)

--- End quote ---
I think JSD was just something they cooked up when the business from JSP was tailing off. The essence of these things is to be profitable, not useful.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod