Which is a key problem - tools acquire more and more "bits", but they typically don't evolve in sophistication, in terms of how they are presented to the user, as they evolve in complexity. Hence, all the really "clever" stuff is often almost unusable, because it requires so much input from the user, the latter has to become an expert in using the program to get the good bits to work ... the tool becomes "dumber", as far as the user is concerned, the more fancy it gets ...
That higher level methods of specifying what a program is required to do have not evolved - virtually all the languages are still caught up in requiring each tiny, boring step to be specified - and that it has to be 100% correctly stated for the end result to work.
Think of it this way: I want my kitchen to make a chocolate cake,; I want the kitchen to come back to me and say, OK, do you want a chocolate cake exactly like I made once in the past, or do you want one that is a variation, or do you want to try a completely different recipe. Okay, a variation today... now what should I vary, and by roughly how much ...
The reality is, still, that I largely have to say, Go to the pantry, find some flour, bring the flour to the kitchen bench; now go find a mixing container, etc, etc, etc, etc .... geddit??
Yes, some of the straightforward, interface stuff can be easily determined these days ... but the gutsy, behind the scenes stuff - the "meat" - still has to be done in the same, laborious manner ... it wore me out, doing it !!!!
As a counter-example, there is a single high level tool for setting up a business system - where all one has to do is specify the Business Rules, in a human-friendly way, for a fully functional system to be spat out ... one and one only, that I'm aware of ...
No I don't 'geddit'. You talk in abstractions that are so far removed from everyday engineering realities as to make them worthless. Fortunately we live in a universe that requires we do A then B to get to C. I am afraid your utopia is still some way off . . . In the meantime, thank you Dennis Ritchie and the host of computer scientists that came along after him to make coding easier and more intuitive.
Sorry Frank, I don't share your views on this one.
Last edited:
Yes, really good software would let you disable parts to customize it for the user. Unless you use a program frequently, you can't remember
most of the "features" that make it difficult to do the basics. Altium is way too complex for simple boards. Programming tools are a different matter.
Last edited:
Well, as a developer, i have to defend Frank.
Pascal was yet an object oriented language, certainly a more structured serious and bug free one than C++.
Web programming ? Script languages (PHP, Javascript etc.), interpreted (slow), and used mainly by the young generation under huge libraries, an atomic bomb to kill single flies, most of the time. Oh, we could dream of something more efficient and fast.
It is very rare to see, nowadays, a web site with no HTML errors, in accordance with the W3C requirements.
Worse than this, 80% of the code of blogs and CMS (heavier and heavier) is not anymore for the content, but for the look ands design.
Javascript ? The worse language a human can imagine.
And, can-you imagine, to code a forum, by example, you have to deal with at least 4 different languages, all together: HTML, PHP, Javascript, MYSQL. Each one with a very different approach, and syntaxe. Not really a good situation, on my opinion.
And because all this stuff is so heavy, slow to load and execute both server side and client side, we have to use caching systems everywhere.
Pascal was yet an object oriented language, certainly a more structured serious and bug free one than C++.
Web programming ? Script languages (PHP, Javascript etc.), interpreted (slow), and used mainly by the young generation under huge libraries, an atomic bomb to kill single flies, most of the time. Oh, we could dream of something more efficient and fast.
It is very rare to see, nowadays, a web site with no HTML errors, in accordance with the W3C requirements.
Worse than this, 80% of the code of blogs and CMS (heavier and heavier) is not anymore for the content, but for the look ands design.
Javascript ? The worse language a human can imagine.
And, can-you imagine, to code a forum, by example, you have to deal with at least 4 different languages, all together: HTML, PHP, Javascript, MYSQL. Each one with a very different approach, and syntaxe. Not really a good situation, on my opinion.
And because all this stuff is so heavy, slow to load and execute both server side and client side, we have to use caching systems everywhere.
Last edited:
Later versions of DOS Tango were great in their simplicity. I'd like to have a similar current program, especially for the Mac.
You could actually not waste half your time trying to get things to work.
Rip Van Winkle, you are. The average PC blows the doors off of anything 30yr. old. ALL our layout simulation software runs under Linux on stock PC's.
Unfortunately guaranteed anecdotal sighted listening tests, nothing but smoke on the water (with all due respects).
Preferably not pay for that bloat in the first place. As far as I recall Protel came on two 5.25" floppies. So perhaps 2Mbytes of code? The price if per Mbyte would be really attractive as the full Altium suite runs to hundreds upon hundreds of Mbytes.
It is interesting that you guys are nitpicking a 30 year old design and layout. '-) That was SOTA 30 years ago, not today. Carl now uses a PC with very expensive professional layout software. He probably has 5-10 times more money invested in the software than the hardware, these days.
My experience with computers goes back 51 years. (note my avatar, it could have been me)
THEN, computers were big and very slow. 20 years later, Carl did the Vendetta layouts, and things were still pretty slow, BUT they could do a lot. Carl might have had to wait 1/2 hr for a simple change. Today, things are faster, so what? What matters is making the right decisions regarding design and layout. We came out pretty darn successful with the Vendetta.
Carl says that today we would stick with reinforced Teflon boards for the low level stuff, but we might use FR-4-8 for line level stuff. I will recommend this to Parasound.
My experience with computers goes back 51 years. (note my avatar, it could have been me)
THEN, computers were big and very slow. 20 years later, Carl did the Vendetta layouts, and things were still pretty slow, BUT they could do a lot. Carl might have had to wait 1/2 hr for a simple change. Today, things are faster, so what? What matters is making the right decisions regarding design and layout. We came out pretty darn successful with the Vendetta.
Carl says that today we would stick with reinforced Teflon boards for the low level stuff, but we might use FR-4-8 for line level stuff. I will recommend this to Parasound.
I was involved in designing and user interface development of PCB layout software in the early 90-ies. Was very sophisticated for the time.
These days my $ 250 Ares from Labcenter is infinitely better, smarter, more options, more intuitive.
These days PCB layout for lowly audio circuits allow you to do anything you want, tweaking individual pads, traces, polygon planes, trace hugging, necking and teardropping, what have you. Furthermore, tight integration of layout software with schematic drawing insures pretty much faultless PCBs on the first try.
As can be seen on the boards shown by PMA, that layout software was VERY primitive compared to what is available even to amateurs today.
Jan
Christophe,
I feel for you if you are writing code and I do remember even in the 80's all the different languages that were out there for different purposes. At the same time this is not back in the day when my father had to have a coder work for a few years writing code in UNIX to run on a tape drive to do things like billing and such when no software existed that you didn't have to develop yourself.
Computers today, the lowly PC is so much more than an entire refrigerated room full of tape drives and hardware could ever do back in the day. Let's face it most people never needed a PC in the first place and most are fine with a smart phone. At the same time there are some of us who actually use a modern PC at the limits of what they can do. I know my CAD package wouldn't even run on an old computer, the code just wouldn't work with old ram limitations and hard driver speeds and size. How many programs actually are written to use the multi-threading and multi-core architecture of a modern processor? I know my CAD uses that, just try to do a rendering on an older machine and compare the time it take to do one picture, I can't wait to get a newer machine with more cores and threads it will save me so much time waiting for the computer to get a job done.
It is on the software developers to use the best language to get the job done, but anyone who has ever written any kind of logic diagram or flow chart knows some people can do this with fewer steps and cleaner execution, it is an art, software bloat is just lazy code writing, you may end up with the same end result but the hoops and loops you have to go through to get there can be mighty convoluted. It still comes down to yes/no ladder logic to come to a conclusion, we all do it every day.
I feel for you if you are writing code and I do remember even in the 80's all the different languages that were out there for different purposes. At the same time this is not back in the day when my father had to have a coder work for a few years writing code in UNIX to run on a tape drive to do things like billing and such when no software existed that you didn't have to develop yourself.
Computers today, the lowly PC is so much more than an entire refrigerated room full of tape drives and hardware could ever do back in the day. Let's face it most people never needed a PC in the first place and most are fine with a smart phone. At the same time there are some of us who actually use a modern PC at the limits of what they can do. I know my CAD package wouldn't even run on an old computer, the code just wouldn't work with old ram limitations and hard driver speeds and size. How many programs actually are written to use the multi-threading and multi-core architecture of a modern processor? I know my CAD uses that, just try to do a rendering on an older machine and compare the time it take to do one picture, I can't wait to get a newer machine with more cores and threads it will save me so much time waiting for the computer to get a job done.
It is on the software developers to use the best language to get the job done, but anyone who has ever written any kind of logic diagram or flow chart knows some people can do this with fewer steps and cleaner execution, it is an art, software bloat is just lazy code writing, you may end up with the same end result but the hoops and loops you have to go through to get there can be mighty convoluted. It still comes down to yes/no ladder logic to come to a conclusion, we all do it every day.
Jan,
I can describe the way we designed even quite complex and sophisticated multilayer PCB boards in the first half of the eighties (not the simple ones as shown here).
1. step was to prepare large (4:1, e.g.) draft on a special raster paper. This paper might had added aluminum foil inside in order to keep precise dimensions, in case the drawing was used for direct photo and film production. For a digitizer, standard raster paper was enough.
2. step was to use a pointer device, similar to PC mouse, with a window and small cross inside to click on every angle break on the pcb track. These break points were scanned by a digitizer and stored in a memory of a mainframe computer, together with info about trace width and hole diameter and pad diameter.
3. step - output from the mainframe computer was a punched paper tape.
4. step - punched paper tape as an input for drill and photoplotter.
I can describe the way we designed even quite complex and sophisticated multilayer PCB boards in the first half of the eighties (not the simple ones as shown here).
1. step was to prepare large (4:1, e.g.) draft on a special raster paper. This paper might had added aluminum foil inside in order to keep precise dimensions, in case the drawing was used for direct photo and film production. For a digitizer, standard raster paper was enough.
2. step was to use a pointer device, similar to PC mouse, with a window and small cross inside to click on every angle break on the pcb track. These break points were scanned by a digitizer and stored in a memory of a mainframe computer, together with info about trace width and hole diameter and pad diameter.
3. step - output from the mainframe computer was a punched paper tape.
4. step - punched paper tape as an input for drill and photoplotter.
Last edited:
Fair enough ... I come from the angle of being a problem solver, and the process of creating 'correct', and appropriate software is very interesting to me - I want to see constant evolution of that occurring, and the forward thrust of the momentum of making it happen began to be lost in the 90's - pretty frustrating to see.No I don't 'geddit'. You talk in abstractions that are so far removed from everyday engineering realities as to make them worthless. Fortunately we live in a universe that requires we do A then B to get to C. I am afraid your utopia is still some way off . . . In the meantime, thank you Dennis Ritchie and the host of computer scientists that came along after him to make coding easier and more intuitive.
Sorry Frank, I don't share your views on this one.
Yes, reason why very few people, nowadays, take care of resources when it is about PCs softwares.
Look at Firefox. Can-you believe it takes >10 second to be launched on my 4 cores 5GHz computer with enough ram to record all the humanity knowledge in text mode ?
It seems that each time the hardware improve, the software is going one step back in efficiency in the same time.
In a way, it can be compared to HIFI. Some designers just add slices of complexity, Diamonds, super pairs, drivers, cascode at each stages in order to get beautiful distortion numbers. Not surprising most of the little details of music are lost in the labyrinth ;-)
Happily, in hifi, some persists in optimizing simple schematics, in software, the smartphone's hardware limitations oblige the developers to get back to good programming practices: "Simple is beautiful" where some intelligence can save sometimes hundreds of code's lines.
Remember the Voyager computer: General Electric 18-bit TTL machine with a bit-serial, single register accumulator and bit-serial access to plated-wire RAM (4096 words). 25,000 instructions per second. To reach the limits of the known universe.
Personally, I think it's a rather nice touch to introduce the possibility of unintended parasitic behaviours,
What are you referring to?
- Status
- Not open for further replies.