There is a fun and still growing area, with a bright future, where you may be able to get your foot in the door, maybe you could specialize in building good, usable web UI's for embedded devices.
I've thought about that but I hate this idea. Web app interfaces for embedded devices? Doesnt that make it vulnerable to attack? Seems like a bad direction to take things.
I learned some Unix and was connecting up to friends systems to get and send emails, which at that time were addressed using chains of exclamation points.
Youngsters today can't believe what was involved in that "bang addressing", so-called since "!" was sometimes pronounced "bang".
The key point is that
you (the author) specified the name of
each and every mail-forwarding computer the email addressed passed through on the way to the destination. Get one wrong, or if a machine was down, and the email bounced back in your face. That lead to many "interesting" conversations such as "have you tried going via ucbvax; that's pretty well connected and might know about kremvax".
Sometimes the past
wasn't so golden!
The "Bang path" it was called..
At various times I was connected via UCSF and also a number of local systems which were run by interesting, colorful people.
I used a huge Telebit Trailblazer modem.. no other modem would do. Anybody remember them? While negotiating the link, they made a unique sound ...
By the early 90s this era was ending as corporate ISPs came onto the scene..
But still, for a while, there was a small local ISP, "the Little Garden" - That ended around 1996.
At the beginning the speeds were very limited for many people.
Nobody believes me when I tell them this but for a long while, all of Wired magazine's net connectivity was through a single Zyxel modem.
Not like today.
But I had a really fast connection at work. Which was fun. Work was fun, imagine that.
Everything was new and changing very very quickly.
Cost.. switches and displays are expensive.. silicon and radio waves and web GUIs are cheap. And extendable/flexible.
Just keep embedded HW on their own net without ANY connection to the wider internet.
There is a fun and still growing area, with a bright future, where you may be able to get your foot in the door, maybe you could specialize in building good, usable web UI's for embedded devices.
I've thought about that but I hate this idea. Web app interfaces for embedded devices? Doesnt that make it vulnerable to attack? Seems like a bad direction to take things.
I learned some Unix and was connecting up to friends systems to get and send emails, which at that time were addressed using chains of exclamation points.
Youngsters today can't believe what was involved in that "bang addressing", so-called since "!" was sometimes pronounced "bang".
The key point is that you (the author) specified the name of each and every mail-forwarding computer the email addressed passed through on the way to the destination. Get one wrong, or if a machine was down, and the email bounced back in your face. That lead to many "interesting" conversations such as "have you tried going via ucbvax; that's pretty well connected and might know about kremvax".
Sometimes the past wasn't so golden!
Ugh I’d forgotten UUCP. Thanks for all the bad memories
It’s also take you a few days before you got the bounces because it was all dialups and they only dialled out off peak when I was working for cheap arses. I do miss usenet before it turned to shit
I regular get asked why I refuse to learn AngularJS. Because I’ve seen things you will never see!
By the early 90s this era was ending as "real" ISPs came onto the scene..
And with them came eternal September and the green card spam.
But still, for a while, there was a small local ISP, "the Little Garden" Nobody believes me when I tell them this but for a long while, all of Wired magazine's net connectivity was through a single Zyxel modem. Not like today. But I had a really fast connection at work. Which was fun. Work was fun, imagine that. Everything was new and changing very very quickly.
I remember when there was a single web page where people could announce new pages, at cern.ch of course.
Then came altavista and yahoo, which were obviously un-scalable. Then google.edu appeared.
Cost.. switches and displays are expensive.. silicon and radio waves and web GUIs are cheap. And extendable/flexible.
Just keep embedded HW on their own net without ANY connection to the wider internet.
There is a fun and still growing area, with a bright future, where you may be able to get your foot in the door, maybe you could specialize in building good, usable web UI's for embedded devices.
I've thought about that but I hate this idea. Web app interfaces for embedded devices? Doesnt that make it vulnerable to attack? Seems like a bad direction to take things.
How would you even do that? If its connected to the LAN and the LAN is connected to the WAN, it is connected to the internet.
I used a huge Telebit Trailblazer modem.. no other modem would do. Anybody remember them? While negotiating the link, they made a unique sound ...
They were quite a pioneering use of OFDM type techniques. I'm not clear why nothing else in the phone line modem world went in that direction, as they worked very well (ADSL doesn't count. That bypasses the telephone network, and just reuses the last mile pair with an OFDM type of signal).
How would you even do that? If its connected to the LAN and the LAN is connected to the WAN, it is connected to the internet.
There is nothing fundamental that says WAN has to equal internet....
Very common on broadcast sites where you sometimes have huge, many Gb/s realtime video multicast on one network complete with 40/100Gb ports, with IGMP and all that pain, then you have a control network at 100 or 1Gb/s, then you have the office lan which might have some very restricted access to the control lan but will usually have no connectivity to the video transport network at all (And there has been serious experimentation with distributing this stuff across multiple sites around the country).
Interestingly that whole world lives on IP v4 and there is no real pressure for v6.
Regards, Dan.
How would you even do that? If its connected to the LAN and the LAN is connected to the WAN, it is connected to the internet.
There is nothing fundamental that says WAN has to equal internet....
Very common on broadcast sites where you sometimes have huge, many Gb/s realtime video multicast on one network complete with 40/100Gb ports, with IGMP and all that pain, then you have a control network at 100 or 1Gb/s, then you have the office lan which might have some very restricted access to the control lan but will usually have no connectivity to the video transport network at all (And there has been serious experimentation with distributing this stuff across multiple sites around the country).
Interestingly that whole world lives on IP v4 and there is no real pressure for v6.
Regards, Dan.
That's a bit exotic, but most large multi-national companies have their own global networks, with just a couple of tightly locked down gateways to the public internet, allowing only e-mail and tightly controlled browser data though.
Simply never connect it to the Internet or to anything that is.
If you need multiple subnets use routing by IP address + router. No DNS required.
How would you even do that? If its connected to the LAN and the LAN is connected to the WAN, it is connected to the internet.
There is nothing fundamental that says WAN has to equal internet....
Very common on broadcast sites where you sometimes have huge, many Gb/s realtime video multicast on one network complete with 40/100Gb ports, with IGMP and all that pain, then you have a control network at 100 or 1Gb/s, then you have the office lan which might have some very restricted access to the control lan but will usually have no connectivity to the video transport network at all (And there has been serious experimentation with distributing this stuff across multiple sites around the country).
Interestingly that whole world lives on IP v4 and there is no real pressure for v6.
Regards, Dan.
That's a bit exotic, but most large multi-national companies have their own global networks, with just a couple of tightly locked down gateways to the public internet, allowing only e-mail and tightly controlled browser data though.
Yep, not the smartest of ideas though and a heck of a lot more than email and browser data is allowed through. Firewall pinholes do require approval of course and get reviewed periodically. There will need to be more like half a dozen or more gateways to the internet as well - if you don't want the peasants to rise up - because latency ya know.
Makes for a fun time when some random other country gets infected with something and it spreads worldwide over the internal WAN. This silly idea of 'perimeter defence' is borken when the perimeter is so incredibly large.
In my experience, the closer you are to head office the more 'exceptions' to the rules there are, as they have never been forced to clean up their historical crap (too big to migrate blah blah). The remotest parts of the world are forced to discard the past and migrate to whatever the latest flavour of the month is, and so in doing so are kept up to date and relatively clean and tidy.
There is a fun and still growing area, with a bright future, where you may be able to get your foot in the door, maybe you could specialize in building good, usable web UI's for embedded devices.
I've thought about that but I hate this idea. Web app interfaces for embedded devices? Doesnt that make it vulnerable to attack? Seems like a bad direction to take things.
Not any more or any less than having any other network interface on the instrument.
How many instruments have poorly implemented LXi interfaces or half-baked command interpreters where you can just telnet in without a password (or with a trivial, well known one that cannot be changed), accept firmware updates from bogus sources, etc. ?
Having an embedded web server on the device that serves you a control panel or something when you open it in the browser if not fundamentally more dangerous by itself. It just needs to be implemented properly and not exposed to the internet.
The cost constraints on many products and the sophisticated UIs and APIs which can be implemented in HTTP and HTTP-like protocols via web UIs and XML make it the obvious choice for hardware products which need extensibility combined with low cost. As I said before, the hardware which would be necessary to implement all that functionality without using a web UI is much more expensive. Why reinvent the wheel?
This is why the next big revolution in the workplace will be machines talking directly to other machines via text-based protocols- but usually without people in between them as is the case today.
Every ipv6 capable device has its own globally unique network address.. Think about that for a minute.
------
You know, there is a body of early research in what is called "ubiquitous computing" (much of it was doe at Xerox's Palo Alto Research Center or "PARC") that you might want to check out.
As it was a vision of a kind of high-productivity- high-functioning cooperative workspace which for some reason we still have not managed to ever replicate in corporate America, but which does often happen in pure science research/academia.
Since you lack exposure to academia, it would also be an opportunity for you to build up some of the knowledge acquisition skills you would otherwise get doing college level research and writing papers.
You might want to check out the Linux operating system because of it's wealth of tools for science. I can tell you that for large segments of the computing world, Linux and other Unixes are the default operating systems for many reasons. Stability is just one of them. Another is Unix's modular nature lends itself to scripting, connecting the (standardized) text-based output of one application to the input of another in chains.
If there is an exception (error), its also handled in a standardized way.
Linux is free so all you need is time and some hardware or disk space you can devote to the task. You likely already have a quite capable Linux box if you have a Raspberry Pi or similar SBC.
The recent RPIs actually are more powerful real computers than many high end boxes were back in the day.
There is really a lot of software available for them, including (for free!) Mathematica.
[Note: If you use an RPI as your Linux desktop, you should use a real hard drive.. and get in the habit of backing it up (or use software RAID for redundancy) because flash memory wears out eventually. Especially you should not compile large software packages on a flash card. That might toast it very quickly.]
Just in very practical terms what that means is that exactly like cheap embedded computers are replacing very sophisticated electromechanical devices in all kinds of areas very rapidly, similarly cheap 'tablet computer" technologies (i.e. touchscreens affixed permanently to the device and APIs accessible via the device's internal network, or a non-routable IP based intranet of things.) can replace almost any control panel for those very expensive devices relatively effortlessly, and in most cases do a far better job. (provide the optimal level of control granularity instead of being constrained by parts cost to only a few control modes.. or whatever)
Because of the economics of scale that brings the cost down substantially.
Designing those interfaces optimally is a difficult task, that's professionally challenging.
It's my understanding that some (Telebit Trailblazers, and similar) are still in use in places like Africa and South and Central Asia where old telephone hardware makes up the networking infrastructure, nothing works as well under degraded analog telephone conditions, over old copper telephone lines and through the old style switching systems.
They were quite a pioneering use of OFDM type techniques. I'm not clear why nothing else in the phone line modem world went in that direction, as they worked very well (ADSL doesn't count. That bypasses the telephone network, and just reuses the last mile pair with an OFDM type of signal).
I'm a failed PhD geophysics candidate with a BA in literature and an MS in geology. My PhD supervisor made me slave labor admin for a MicroVAX II. My sole formal education in computing is 1 credit hour of WATFIV taken during my undergraduate work.
All I know about computing came from reading books and talking to people. If someone mentioned a book as being good I'd buy it and read it. The computer section of my personal library consumes 80 ft of shelf space. The opposite side is geosciences, DSP and general math.
Higher education has become a scam, so you are right to be wary. Rather than get a degree, take calculus and physics on a relaxed course load where you have time to work ALL the problems, not just the assigned problems. Don't take physics I until you've finished calculus I.
Get a copy of the C standard and read it. Pay particular attention to where things are undefined or implementation dependent. Make sure you NEVER do those things. Learn forth. Mecrisp is very good.
Get a copy of the "Art of Electronics" 3rd ed by Horowitz & Hill and read it cover to cover twice. If you find it heavy going get the previous edition of "Electronic Principles" by Malvino & Bates The current edition is $200 which is ludicrous. An older edition is fine. H & H will get you up to date. Build the circuits and test them. Document it all in a notebook. Build the interesting circuits dead bug style so that you have trinkets to show.
Get a copy of "Advanced Programming in the Programming Environment" by Stevens & Rago and the "Unix Network Programming" by Stevens. I've made a lot of money from the examples in those books.
Do a few fairly sophisticated projects, e.g. sub-centimeter differential GPS using uBlox modules and ESP32s. Keep a log of your work either online, in a notebook or both. Such a unit would be a very attractive gadget in the real estate sales market for locating property boundaries.
Locate some target employers. Find out where the staff go after hours and start hanging out there when they do. Freely mention whether you're in or out of school at the moment and what you're doing. I dropped out twice before I finished my BA just out of boredom. If you can hold up your end of a technical conversation with useful literature references you'll eventually run across someone who recognizes your abilities. If someone is chatting about a problem and you can point them to the answer, they will get very interested in you. But don't EVER try to fake it. Say what you know and what you don't know.
I met a friend when he appeared in my boss's office to fix her computer. I had an arcane network protocol problem I was grappling with. I asked him the question. He said he didn't know, but that when it did A it did B. I said well when it does B it has to do C. In about 5 minutes we had answered my question. He has no college degree. He went through an electronics trade school course. He's now a senior manager for AT&T consulting.
HTH
Have Fun!
Reg
I'm a failed PhD geophysics candidate with a BA in literature and an MS in geology. My PhD supervisor made me slave labor admin for a MicroVAX II. My sole formal education in computing is 1 credit hour of WATFIV taken during my undergraduate work.
All I know about computing came from reading books and talking to people. If someone mentioned a book as being good I'd buy it and read it. The computer section of my personal library consumes 80 ft of shelf space. The opposite side is geosciences, DSP and general math.
Higher education has become a scam, so you are right to be wary. Rather than get a degree, take calculus and physics on a relaxed course load where you have time to work ALL the problems, not just the assigned problems. Don't take physics I until you've finished calculus I.
Get a copy of the C standard and read it. Pay particular attention to where things are undefined or implementation dependent. Make sure you NEVER do those things. Learn forth. Mecrisp is very good.
Get a copy of the "Art of Electronics" 3rd ed by Horowitz & Hill and read it cover to cover twice. If you find it heavy going get the previous edition of "Electronic Principles" by Malvino & Bates The current edition is $200 which is ludicrous. An older edition is fine. H & H will get you up to date. Build the circuits and test them. Document it all in a notebook. Build the interesting circuits dead bug style so that you have trinkets to show.
Get a copy of "Advanced Programming in the Programming Environment" by Stevens & Rago and the "Unix Network Programming" by Stevens. I've made a lot of money from the examples in those books.
Do a few fairly sophisticated projects, e.g. sub-centimeter differential GPS using uBlox modules and ESP32s. Keep a log of your work either online, in a notebook or both. Such a unit would be a very attractive gadget in the real estate sales market for locating property boundaries.
Locate some target employers. Find out where the staff go after hours and start hanging out there when they do. Freely mention whether you're in or out of school at the moment and what you're doing. I dropped out twice before I finished my BA just out of boredom. If you can hold up your end of a technical conversation with useful literature references you'll eventually run across someone who recognizes your abilities. If someone is chatting about a problem and you can point them to the answer, they will get very interested in you. But don't EVER try to fake it. Say what you know and what you don't know.
I met a friend when he appeared in my boss's office to fix her computer. I had an arcane network protocol problem I was grappling with. I asked him the question. He said he didn't know, but that when it did A it did B. I said well when it does B it has to do C. In about 5 minutes we had answered my question. He has no college degree. He went through an electronics trade school course. He's now a senior manager for AT&T consulting.
HTH
Have Fun!
Reg
Thank you, that is very encouraging.
My new concern is the demand free freelance engineers. What type of client hires a freelancer? In what situations?
My new concern is the demand free freelance engineers. What type of client hires a freelancer? In what situations?
You should note that being a freelancer has many attributes in common with one-man businesses.
You might like to consider whether or not you know enough about
business to become a freelancer, in particular how businesses advertise, select work, ensure that the business relationship doesn't become sour, and avoid litigation.
Not impossible, of course, but as least as difficult as the technical aspects of freelancing work.
My new concern is the demand free freelance engineers. What type of client hires a freelancer? In what situations?
You should note that being a freelancer has many attributes in common with one-man businesses.
You might like to consider whether or not you know enough about business to become a freelancer, in particular how businesses advertise, select work, ensure that the business relationship doesn't become sour, and avoid litigation.
Not impossible, of course, but as least as difficult as the technical aspects of freelancing work.
I do freelance marketing right now, so I am familiar with those topics. But I'm trying to see which situations a business would hire a freelancer vs hiring a fulltime employee. Some have said that they do this if a project is behind schedule or their full time staff doesnt want to do it - this is not something I would want. How often are freelancers used for fresh projects?
Companies hire freelancers to do things their regular staff don't have the skills to do.
When I was freelancing I was hired by word of mouth. I had long term business relationships with people in a number of organizations and companies and they kept me busy.
This is the kind of project I did: My goal believe it or not was one where you actually try to automate yourself out of a job. Do a project so well and with so much attention to how it is going to be used that it almost runs itself.
Make good choices at the outset on platform and document the things you do well.
Paradoxically, that is how to get lots of work!
My new concern is the demand free freelance engineers. What type of client hires a freelancer? In what situations?
You should note that being a freelancer has many attributes in common with one-man businesses.
You might like to consider whether or not you know enough about business to become a freelancer, in particular how businesses advertise, select work, ensure that the business relationship doesn't become sour, and avoid litigation.
Not impossible, of course, but as least as difficult as the technical aspects of freelancing work.
I do freelance marketing right now, so I am familiar with those topics. But I'm trying to see which situations a business would hire a freelancer vs hiring a fulltime employee. Some have said that they do this if a project is behind schedule or their full time staff doesnt want to do it - this is not something I would want. How often are freelancers used for fresh projects?
Given that marketing is all about understanding what the market wants, I can't quite reconcile all your statements.
Anyway, a business takes on a freelancer for their specialist experience not possessed in house (optionally transferring that knowledge to employees), or because the specialist experience is not required for long, or because they need another body for a short period, or because they want to skirt around some employment laws.
As to "how often", a correct but unhelpful answer is "16782 times". Answering questions is easy; finding the correct question to ask and answer is more difficult.
Reconcile my statements? Its called market research! Asking those on the inside questions about the market.
Given that marketing is all about understanding what the market wants, I can't quite reconcile all your statements.
No, marketing is making people believe they need your product!
BTW I spent most of my career as a contractor. Typically about 3 years each place except for the last which ran 8 years. I did not look for work. It found me. At the start I worked through a one man job shop for about 4 years until I discovered my SEP funds had been embezzled. I did get the money back, but it took a while.
I've been laid off twice. The first time was my last regular employee job. I bought a Sun 3/60 and taught myself Unix system admin and lex & yacc over the course of 5 months. My first contract task was a perfect fit for lex & yacc. The client had allotted 2 months for completion. I finished in 2 weeks along with doing a lot of other stuff on the side. I had worked 10-12 hours a day during the 5 months. Pure luck, but it had an excellent payoff. It established my personal brand. At one point I had 2 regular employees working exclusively for me for a couple of months!
Throughout my career I spent a lot of personal time and money buying and reading books I thought might be useful in my work. This gave the perception that I worked much longer hours than I did. My hourly rate made up for only working 35-37 hours a week. My Dad was a professional engineer, but I grew up in a restaurant because he got tired of engineering sales. It was very valuable experience as it taught me good customer service skills. If I finished a task close to quitting time, I went home, I did not run out the clock, nor did I bill my client for eating lunch as I've seen other contractors do all too often.
The 2nd time I bought $15k of computer gear and started developing some seismic processing software. I went to walk the dog during a test run and when I got back there was a message from a friend informing me that an old boss was looking for a "Reg type". That gig lasted 6 years before the merger and 2 years after even though the person who hired me had moved to another assignment. I put $100k of labor into that software and never made a dime from it. But I learned a great deal. I'd do it all over with no changes.
If you don't have a paying job, you should be doing real work on your nickel. Stuff that is important to potential employers. If you won't invest in yourself, why should anyone else? Attitude makes all the difference in the world.