How many programming languages do you know?

I know C, Java, Python, BASIC, SQL and HTML, Bash and Batch(Windows/DOS).

 

Indeed ...

Additionally ...

What you say may have been true at one time, but today, on modern hardware, it is certainly not.

Even if your assumption ( which I must assume it to be ) was correct, the way an interpreted language and the way a compiled language are executed are a world apart, perceived optimizations you can make at the C level and lower mean nothing while executing PHP ... you should know this ... apparently ....

I don't deny a little knowledge can go a long way, so long as that knowledge is correct, and applicable.

 

Well that's a new one on me -- it didn't used to optimize down the same. I guess the people behind GCC are getting sick of being called a third rate compiler compared to even two decade old Borland ones.

When did that change occur? I know about six or seven years ago there would have been a pretty big code difference between the two... The output from turbo C was similar, and older C's were notorious for it. There IS a measurable code size difference in WinAVR and the Arduino compilers, which I believe are just over-glorified wrappers for GCC. Wonder if that persists across processor targets...

Though as x86 assembly it's still **** -- since it's abusing ebx for ecx's job. WAIT...what the devil processor target is that? Oh I see, AMD64/EM64T/Intel 64/whatever they want to call it this week. Cute. Kinda queer that it pushes/pops RBX without actually using it and no real need to preserve it... wastes a long addition on what should be an increment... Though that's the hallmarks of a cross processor compiler using motorola style notation. They made it smart enough to make both ways basically the same code, but not smart enough to do a two byte "inc ebx" instead of the 4 byte "add ebx,1" -- guess that goes hand in hand with the AT&T/Motorola style notation of making the assembler too stupid to recognize the data size by register name, so you have to throw in all those extra $ and % for no good reason. (What is this, a 6502?)... but I guess that goes with the whole cross processor thing where Non-intel legacy has 'addressing modes' on the value and Intel legacy does it by the register.

What's it do on a 32 bit target? I might have to run some tests to see where things are currently at -- which is the cautionary lesson here. Compilers like GCC are moving targets -- what was true of them a week ago could be fiction today -- much less what several years development can lead to.

 

Yours is a habit of the industry: 6 or 7 years is a long time to call something knowledge, at that time it took a celeron around 1 hour to compile PHP with a standard set of extensions, today on a core i7 the same task takes literally 7 seconds flat to the generation of php-cgi. The faster we develop the faster knowledge becomes useless ... take how people saw java 7 years ago; something that was "enterprise", "huge", complicated, a resource hog and something that anyone with sense should stay away from if they didn't have the budget for enterprise hardware. Today Java is holding up the biggest mobile enterprise on the face of the earth, on the most advanced embedded hardware the world has ever seen. It's habit to retain knowledge for longer than it is useful because the logistics of keeping up to date knowledge on everything you learn over the years is that in itself is a full time job, which nobody will pay you to do. As everything does, it comes down to money ...

You obviously know what you're talking about, moreso than me, but I suggest that while having that knowledge is valuable, it shouldn't take very long for you to gain the same applicable knowledge in the frameworks you actually execute, because your assumption that this knowledge can pay dividends has never been, to my knowledge true, at least for PHP. Interpreted languages just don't work like that. You would be far better equipped if you spent some time doing that, I don't see that it would be too much trouble for you, at all. With your knowledge of the low level you'd be a valuable member of the team working on PHP and it's extensions, I'm sure we'd all enjoy your input. Don't sit and complain that there are flaws, sit and fix them

I believe AVR is a subset of 32bit instructions with some special instructions for ARM, this is frozen in time, I don't expect it will develop, and am not surprised that it's far behind ... the work around is multi-core, this is industry standard, being that the physical limits of current hardware design have already been reached, adding cores is the only way to move forward. All of the programming of the main contenders in the embedded market is Java (droid) and C# (MS/iOS) or Objective C (ios), so I've spent no time writing C for these targets, I couldn't tell you if they have been optimized, any my description of what AVR is is probably wrong ... but being that it took so long to optimize the hardware we are discussing it's barely a surprise. Again, with your in depth knowledge, why not help out in improving those platforms ?

I have a cross compiler for arm installed, but as I said, it's pointless to look for any real optimization here ...

You don't have to answer if you don't want to, but what is it that you are currently doing for a job, and what afforded you this ( impressive ) knowledge of assembly ? Are you involved in any opensource online projects, and would you get involved in any ?

 

Your post reminds me of something I'm always telling people -- "3 years is obsolete, 5 years is the scrap heap" -- and I fell into my own trap

More frozen in time than you think -- AVR/Atmega is a 8 bit target... and there are constant improvements to it because it is a simple small low power device easily embedded into many projects... though the API has been ported recently to ARM targets as the price of things like ARM3's is dropping into spitting distance from AT128's. See the new "teensy 3.0" which goes from the 16mhz 8 bit AVR ATMEGA32U4 used in the 2.0 to a whopping 48mhz ARM M4.

A real laugh being that even 8 bit AVR's are more capable of floating point math than most ARM chips -- hell, even with the so called 'vector floating point unit' on them ARM processors aren't as capable per clock at floating point math as a 8087 co-processor circa 1980! I only in the past two years started programming for ARM, and was SHOCKED to find out just how pathetically crippled they are at performing real math. (admittedly, having a 80 bit FPU on a 8 bit data path next to a 16 bit CPU was WAY ahead of it's time)

I've used the Teensy++ 2.0 (AT90USB1286 powered) for a number of projects due to it's simplicity and low low LOW power... and it's compatibility with other prototype systems like the Arduino. I really prefer the Teensy's over the Arduino since they're easy to fit into a socket on a board, they have REAL USB built in, and they plug right into breadboards... things that to be frank the Arduino should have offered since day one.

Assuming you need that type of power for a project -- something I've learned after many MANY years in the industry is that if you can get away with cheaper hardware, DO IT. There's this noodle-doodle concept of just throwing more hardware at bad code, and the whole industry is suffering because of it.

I have no secrets anymore.

Retired due to worsening health at the moment -- actually spent the past seven years weaning clients off my support since I'm actually on doctors orders not to work -- which is why I run a bicycle repair shop out of my garage during the warm seasons, still build systems via word of mouth at a flat rate, go out busking (street performing) on my EWI on summer weekends, maintain a number of websites I've ghost written, and make a royal pain in the ass out of myself by trying to help others for free on forums. I have difficulty not working and just sitting around all day with my thumb up my arse.

Being retired has let me put time into my own projects, which has been nice as I've been learning far more than I did when I was at the whims of others. Given my last work experience involved my screaming at the employer "You hired me for my knowledge, then ignored EVERYTHING I tried to tell you, then blame me for your incompetence. You think you can do better, you do it, otherwise WHAT THE **** DID YOU HIRE ME FOR!?!" -- even more fun when three months later they come back hat in hand apologizing and begging for help, and you are financially at a place where you can afford to tell them exactly where to stick it.

My first computer was a Cosmac Elf I built in the basement in 1977. As such I cut my teeth on RCA 1802 machine language I had to HAND COMPILE myself and enter eight bits at a time on a row of toggle switches. Shortly thereafter I moved to a TRS-80 Model 1, and then had access to a PDP-11 ( and later a DEC Rainbow) at my old man's work. I did a lot of Z80 and 6809 (on the Coco) assembly as a teen. By the early '90's I spent a good four or five years writing accounting software using Clipper and Paradox, but then changed gears, opened up my own mom & pop whitebox shop, and backed away from programming for a while... then I came back at the start of the last decade to do websites -- and found the efforts of others... lacking. I ended up teaming up with Dan Schulz (deceased) for a number of years on web dev projects -- I taught him the common sense concepts of programming, he came back at me a year or two later teaching me separation of presentation from content and semantics. The apprentice became the master - as it should be.

Basically I've got some three and a half decades programming experience on everything from 8 bit to 64 bit... if it wasn't for my health and mental state, I'd be kicking names and taking ass. Especially with my having worked for the likes of DEC, Marstek, Kodak, and even a short late '90's jaunt as a Apple service tech adding to my three and a half decades experience.

I even go back and learn old stuff just to know it when I can spare the time. That was the entire reason I made a new pac-man game for 1982 DOS machines a couple years ago, and then gave myself a crash course in 6502 machine language to port it to the C64 in November. NOT that the C64 version can really be called a port with the new graphics, and that it's written in 6502 machine language and C, instead of the x86 machine language and Pascal of it's DOS counterpart. Prior to October I really had never done any serious 6502 coding, despite having been knee deep in the 8 bit computing scene of the late '70's and early '80's... Likely because I never once saw an Apple II in real life until well after the PC/x86 was an established force. But sure, they were the driving force of the computer revolution -- their revisionist history says so!

My AVR knowledge is also recent... started out with an arduino, but the platform was difficult to interface to hardware wise and it lacked enough pinouts for my taste -- which is why the teensy++ is my weapon of choice for one of my other current projects, the Zoot -- which I intend to be an open hardware and software platform equivalent to wind controllers like the Yamaha WX series or the EWI USB. (and yes, EWIUSB.com is one of my websites). It's got capacitive keys, a 1.5" OLED display, and a Playstation 2 joystick on the bottom for use in controlling pitch bending and other MIDI effects... outputs MIDI over USB, is powered off the USB port. Hardware on the first prototype is 99% complete, software has been sitting at about 70% complete since October. Holidays and health put it on the back burner, but I want to have it done by march so I can take it out busking in the spring. 2nd prototype will likely have it's own battery supply, may even incorporate it's own synthesizer using a Raspberry Pi or Cubieboard, and a real MIDI output so I can use it with older dedicated modules. I still have to investigate the legal aspects, but was thinking that by the time of the third prototype I may look into mass fabrication since I could probably street these for under $100. First prototype's cost was $62 for parts, and that's mostly due to development issues and single order off the shelf parts. If I could order 100, using fabbed boards instead of point to point wiring and machined/molded plastic cases instead of hand carved, stained and polished wood (dangerous for someone with a parkinsonism to be doing) I figured I could make them at $23 a pop (not counting labor).

I am a fan of the open source CONCEPT, but I hate it's naiveté and the noodle-doodle whackjobs at the FSF who are a mix of '80's back room server geeks who never mattered for anything real, and high school/college kids who don't know any better. The worst of the lot are the career educators and professional lecturers who've never had to work an honest day in their life... basically the all the worst parts of "societies" with "manfesto's" and "social agenda's" all rolled into one. When someone starts spouting off about needing to change how people think -- be afraid, be VERY afraid!

My distaste for "open sores" becomes particularly heated when fat bloated idiocy like the GPL is applied to projects that have no cost applied to them. When you have to use an over-glorified EULA filled with legalese contract law to circumvent real laws, in the name of "freedom", well... someone just might have to explain what the word "freedom" means since their "rah, rah, rage against the machine" cheerleading is so full of socialist rhetoric it ends up one step removed from the world of Orwell's fiction. That said document is larger than the founding document of most world governments, you're BS alarm should be going off full blast!

Which is why like the folks at SQLite I release my works to the public domain... since as you'll find the disclaimer on most of my websites says for **** sake if you're going give something away, JUST GIVE IT AWAY!. See my recent game releases for C64 and DOS which are more open than open source -- being public domain releases with full source code... giving RMS the finger the whole way.

As to my working on public projects with others, I have GREAT difficulty being productive in collaborative environments, and the complete lack of communication that goes with the idiotically useless half assed collaboration tools and software (like say... github) only further frustrates me. In the '90's I wrote entire commercial double entry accounting packages that are still in use today single handed -- As a decade of my report cards from the '80's read "Doesn't work well with others.

Though I may have to call in help on one of my projects just because after talking to certain folks (I call them Python wankers) a project that should have taken me a month has been sitting on the back burner with ZERO activity since June... A sufficient amount of time I might have to start over from scratch just to remember what I've already done. NOT that I'm sure I don't want to just cancel this one NOT because it's not viable, but due to certain people in the Raspberry Pi community making me want to line every last one of them up against a wall fascism style... but then that might change since I'm ditching that hardware target in favor of the cubieboard anyways.

Of course working with others is hard when your gut reaction coming into existing projects is that it all needs to be thrown away and started over from scratch...

Sorry for the long-ish post -- but you asked! ... and as a famous preacher once said: "I'd make shorter sermons, but once I get started I get lazy about finishing."

 

Back to money again, it was never a question of needing multi-core arm devices, but a question of would it sell the next ipile-of-crap. Tegra3 I believe is quad core arm, because it sells, not because it's needed, but because people will pay several thousands of pounds/dollars/chickens/whatever over two years of a mobile telephone contract to have it in their life.

I too love the concept of opensource, I find taking part to be quite stressful, but then if you can you should, and it sounds like you've done more than your fair share already

I too love the concept of teams, but I find like you that not only do I want to scrap everything, that everything needs rewriting once re-implemented by anyone on the team, it's very difficult to find a set of programmers with similar standards, and very difficult for me to conform.

It's interesting reading your posts, we appear to think quite alike, I can only hope in another 15 years I am as knowledgeable as you. On adding hardware, I think of everyone as sheep; the problem has become that it is acceptable to write software that only scales by adding hardware, we live in a "global village" and everything must scale or perish, it's become acceptable because that is how the industry scales, not by improving their platform any, but by simply gluing on more cores.... they are a product of their environment ...

 

Well, you're thinking much bigger/pricier devices with much lower sales counts than I am. I'm thinking things ranging from X-Box/Playstation controllers, mice, bar-code scanners (what I wrote documentation for at Marstek), right down to stupid little things like music/audio playing greeting cards. Microcontrollers for motor control in things like industrial robots also tend to use much simpler chips like AVR 8's. You'd be surprised how many devices from microwaves to timed release air fresheners are built around 6 or 8 pin ATTiny's these days. At as little as 30 cents a pop, AVR's are in all sorts of devices you'd never think even have a micro-controller in them... That ten dollar Walmart coffee maker? Dimes to dollars there's either a PIC or a AVR in there.

... and somebody's gotta program them. There's actually good money in it too if you hook up with the right companies. It's why learning and prototyping platforms like the Arduino exist. Core M4 and M3 SoC's are starting to come down into real world usable prices for things like that, but they're still ten bucks a pop or more -- a low end AVR or PIC still costing a twentieth that. You want to program for these devices - you're either doing it in C++ or Machine language. (or dumbed down C like Arduino)

They are also useful when dealing with larger processors or development systems that don't break out enough pins or devices that just lack enough 'proper' raw inputs. See my biggest complaint about the Raspberry Pi, where the GPIO connector may provide SPI, I2C, etc... but a dozen or so raw digital I/O pins and a few analog to digital pins would make it far, FAR more useful... and the only thing exposed on it for general input is USB, meaning you need something like a AVR acting as a middle-man... (My experience with the Pi left me asking "Just who is this FOR?!?") It's why I am much more enthusiastic about things like the Freescale MK20DX128 based Teensy 3.0 on the low end, and the Allwinner A10 based Cubieboard on the high end as they're both going to have useful pinouts meaning I won't need to hook up an AVR in-between the real world and projects I build on ARM.

I'm also thinking embedded projects -- you make something like a ATM, Cash Register or Lottery machine, are you going to throw away money on a Core 2 or multicore ARM ($40 or more per chip) when an embedded Atom ($27 for a tray of 16), or SoC ARM ($4 to $10/chip) can do the same job? If your selling those systems, you're balancing ease of development vs. cost of hardware.

Hell, when it comes to a typical cash register, there's no legitimate reason to need anything more powerful than a 8mhz 8 bit processor anyways... which is why it's a laugh when companies waste time building full blown modern desktop PC's into them. It's a waste of money that effects your bottom line on every sale.

Basically, the difference between task oriented computing and general purpose computing.

 

8MHz, huh? The Hong Kong Space Museum Planetarium is "automated" with a 1 MHz Z-80. (The tape system [to record and play back whole shows] and the floppy/moving sound controller each had its own processor.) And we did a 400 foot parallel data bus when everyone said it couldn't be done even if you put it into a double-walled shield tube - but we had it running in the powerline trough. Back in 1978.

Today people throw power, money and crappy bloated code at problems - back then we just solved them.

 

Pascal , HTML, C++ and C#. A bit of VB too.

 

I know java, c++,oops and oracle.

 

"Oops" (I think you meant OOP) isn't a language, it's a technique. Oracle isn't a language, it's a company. It sells three main database programs, MySQL, Oracle Database and RDB, none of which is the name of a programming language. (They all use the SQL language for programming.) Oracle also sells (or gives away) other products - like Java, which is a programming language.

 

Just visiting the forums again.

Got some serious old dogs running around in here. I program in c language and basic language variants. Anything there's a syntax manual handy for. My relationship with computers has developed oddly so...

Never had any reason to use an assembler or any of that. I've read a bit , played with the some emulators a while back but I'm never one to do a puzzle for the sake of having it done.

Professional development, I was mostly in QA, configuration management, related areas. That kept me working in higher level languages, interpreted a lot. Shell scripting for whatever someone decided to plug in this week.

At the moment I'm really only actively working around a remote LAMP stack, from my local windows box. It's all textpad and FTP for me...

I wound up writing some interesting stuff but it's best if you can get to a lower level when you need to, and no sooner in my opinion.

Having had to encapsulate a couple of interpreted languages into other interpreted languages, the number of hours I could spend complaining about the kludges would total a lot more than the hours required, in the end, to get what the "star" wants on their desk.

I think the thing about commodity computing is that if I want to write and sell a cash register system, I can get my first 100.000 units to market for less than it will get me to design, implement, UL approve, etc... etc... for a new model of cash register.

It's so much cheaper that you can sell them a new PC when they need to "upgrade their cash registers" because some other aspect of the overall business system changes, even if the OS or hardware in the "cash register" hasn't had a generation yet.

Considering how far computing has moved while most people still want an "appliance model" for their interaction with most electronics it is likely to be a better value, in the end, for the company to replace all of those "cash registers" than it would be to try and have people across the country upgrade the software on them and perform a changeover/training on the new "software in a computer" in any sort of synchronized fashion. People get balky and crap themselves if it acts too much like a "computer".

That's changing, but it's a social issue on a generational scale. And the part about designing, manufacturing, deploying, training, and integrating custom solutions will still persist while the social issues are dealt with.

That particular problem will simply be solved by the disappearance of the "cash register" I think. That checkout line is such a bottleneck, and self-checkout is only half a solution. Don't you think?

Not to derail the issue. It's just the benefits of the leaner, meaner hardware and code are completely overwhelmed by other drawbacks across the business model.

Minimalism requires expertise, and expertise is a NOTORIOUSLY hard to handle commodity.

 

I know c, c++, Advance JAVA, From database side i am strong in ORACLE 10g. For developing a webpage i have knowledge in HTML, JAVA SCRIPT., XML(little bit).

 

PHP, mysql, HTML, CSS, and Joomla 2.5 or 1.5 (CMS).

 

Just Assembler

 

The ones that I can use without referencing much literature are PHP, JavaScript and C++, but most object-oriented programming languages follow the same syntax and I could get by in ASP .NET, for example, if I had to (gun to my head).

PS HTML and CSS are not programming languages.

 

Yes, HTML and CSS it is not programming languages.

I know C++, C#, PHP, Assembler, Delphi.

 

How do you like C#? I wasn't too into the WYSIWYG GUI that they use, guess I'm old fashioned and I'm only 22. Does it save time or have other advantages that C++ doesn't?

 

Programming languages php, C, java, javascript
Design languages HTML, CSS

 

How on earth a single extra comparison in:
for (t=0; t<6; t++)
will make the program 2 x bigger and slower? Did you ever check the ASM output?