I suspect you meant "without any judicial oversight or review." Otherwise, right on, this is much more of a letter that may be read and actually do good. Then again, since it's not printed on hundred dollar bills and since the Senator is twit/stooge enough to make such a foolish comment in the first place (even if under the idea that "there's no such thing as bad publicity"), he's probably solid bone from ear to ear, and since it's Utah, probably a real moron of a Mormon, but I repeat myself.
For the love of all that is holy, we don't need more script kiddies or PHP programmer-wannabes. The article's suggestion of Logo is actually pretty good - instant results; for a clean scripting language, give them Ruby (and perhaps they can graduate to Perl later if necessary e.g. for better library support), but dear God let's not create any more Visual Basic users (bah; I refuse to refer to them as programmers). Even Delphi (or Kylix) would be preferable (for RAD Windows development), or Turbo Pascal in DOS (i.e. a Win9x console), and if they can handle Delphi they can handle Visual C++, especially with all the wizards it has. Preferable not to start kids dependent on Windows, though; the suggestion of the web as a first "platform" is a good idea (especially as it can do graphics relatively easily). My kids are gonna use Unix and they're gonna like it.
The next version of PHP might be able to do the same for Postgres and Oracle databases and they are considering the ability to be able to upload the contents to tables named "bar" as well (but you may have to buy the Zend engine for that).
Amen to everything you said (I may add your comment to my page). See also my PHPSucks page. And fortunately my employers do let me use Perl (and C when I think it's justified).
(Hm, since your nick is "Christianfreak" you probably will know what "amen" means, I guess I can leave it in.:)
Documenting the format is a good step, but if Microsoft changes it regularly, Microsoft will retain the advantage that only their programs are compatible with the "standard" set by Microsoft.
Perhaps a requirement for the company "owning" the standard to release changes a fixed amount of time before they release a product would solve that problem. And also it doesn't do them much good to add to the spec if their current software can't read it; they'd break their own software. Or, if their software can read it as "unhandled feature" and still read the document, other readers will be able to do the same. If not, they will be seen to be in violation of the law and be fined or declared invalid as a government supplier, or at least have their new product declared unusable.
Big business seems fairly safe behind the SEP field it has erected. There are probably 20 comments (at +3, anyway) that say "write your congressman / senator / president / pet hamster / representative" or "start a peaceful protest" but it's always directed to others, with the extremely offensive implication that "I, the commenter, am the great leader, do what I say, to me be the glory, but I am too important to actually go among the rank and file and write a letter myself, I am too busy being the filled-with-holy-zeal visionary." Why aren't there any comments saying "I have already written my representative"? Apathy? How hard is it to get paper, a pen (or printer), an envelope, and a stamp together, look up an address, and put a few words down?
Now for the disclaimer. I haven't written either. That's because I'm in the US on a (TN) visa, and I don't believe it would do much good for a non-citizen to write (and it would be dishonest to pretend, even by omission, that I was a citizen). (My fiancée is a US citizen, and I will probably get citizenship eventually, but it will take at least five years after getting a green card.) I could, I suppose, send money to a representative, but (a) I don't believe that's right any more than it is right for corporations to do it and (b) I'd probably get arrested by Big Brother for government subversion by a foreign national;-).
Remember, there was a time when nobody could imagine the US government waging war against peaceful individuals. Today, the US holds more prisoners per population than any other country in the world -- and the vast majority are drug users or dealers who have initiated no force against anyone.
Yes! Won't somebody think of the poor, peace-loving drug dealers?! OH THE HUMANITY!
Nice troll. Very cute. I guess the reason why you haven't been flamed yet is that the Perl programmers can see it's a troll and the PHP users don't care either way.
No, free(x) doesn't assign NULL to x. There's a good reason for this: avoiding unnecessary work. Yes it's minimal work, and lots of systems will define a macro that will do the free and assign NULL too (ISTR MFC has something like that), but it counts over millions of iterations. C++'s delete operator also doesn't assign NULL, for the same reason. Usually the variable you just freed will go out of scope soon anyway, so why waste time writing to it?
If you've worked with assembler you'll see that the often-made jest of C being portable assembler is quite true - working in assembler long enough will have you wishing for exactly the features that C has (functions with local variables and named parameters, the various looping constructs, expressions). And it goes that far and no further, and doesn't hold your hand. For some people and some projects this is fine. For some it isn't, either because the people aren't skilled enough or the tradeoffs in using a higher level language outweigh any possible advantages.
Which language to use is a combination of project applicability, personal preference and the collective experience and skills of you and (if applicable) your team. Saying "C is bad" or "don't use C because people have made mistakes with it" is narrow-minded and stupid. Use what works for you. Unless it's PHP, in which case you need help, quick.
Obviously you've never held a real camera, like one of the Nikon F-series models. Drop the G3, it looks like little bits of plastic would go skittering everwhere. Drop the F5, look out below.
Oh, and by the way, the article wasn't that funny.
No, PHP was designed (and I use that term loosely) to make dynamic web pages, and the database functions were - like most of PHP - thrown in as an afterthought, unlike perl's DBI, which was designed to be OO and have a consistent interface.
although it can be extended (like Perl) to do a lot more. The routines PHP uses for database management have been thoroughly refined and optimized to achieve maximum speed
That's content-free. All languages can be extended, even Visual Basic, and they all (except perhaps INTERCAL) claim that their libraries have been "thoroughly refined and optimized to achieve maximum speed." Of course, some do better than others (see Doug's shootout).
I believe it is generally accepted that for databasing, using PHP will result in faster, more streamlined, and smaller code.
And I believe the earth is flat, but it isn't. Or, BZZZT, WRONG (zero for three)! And databasing? PHP isn't neither particularly fast nor compact.
Sad but true: PHP sucks. It's a half-assed attempt by some lame twits to make a programming language for non-programmers, which is sort of like trying to make a car for non-drivers. Whether they succeeded or failed at that goal, we all still lose, because of the horrible proliferation of all the crap PHP scripts on the world.
Dijkstra would probably revise his famous quote for PHP's benefit to: It is practically impossible to teach good programming style to students that have had prior exposure to PHP; as potential programmers they are mentally mutilated beyond hope of regeneration.
Locals in Perl, are, for almost any practical purpose globals. I really hope you meant my-variables.
Unfortunately for Perl, its my variables are what are generally referred to as "local" (usually stack-based, with a lifetime extending only to the end of the scope in which they were declared[1]) variables, whereas those created (overridden, actually) by local are just temporary scoped replacements for existing globals. Not too long ago someone completing a pre-interview screening task for us (a supposedly experienced perl programmer) managed to (consistently) use local where he should have used my.
[1] Of course their referants may last longer if a reference is kept in an outer scope.
It is disingenuous in the extreme for them to pretend that they merit the fees they charge because of any sort of "cross-indexing" and "editing". That may be valuable for some, but I'll take the raw text files, thank you, and let grep do the walking (and/or set up my own indexing with e.g. htdig). If they complain about the download costs, another company would probably be glad to host the raw data in exchange for the free advertising they'd get by having their name on the page.
It looks fairly sensible, but the main problem is that (1) good programmers know these things and (2) bad programmers usually can't or won't learn. This makes the audience pretty narrow, i.e., inexperienced programmers with decent raw skills.
And this bit made me laugh (2.5):
What do you do when you start to run out of low-hanging fruit? Well, you can reach higher, or chop the tree down.
It just seems like a funny metaphor (picture it in your head, chopping down the tree is sort of overkill just to get more fruit:), although I understand what he's getting at.
Saying design is not "more fundamental than" debugging is decidedly not saying that design is unimportant, just saying that in the Real World, debugging takes more time and is a skill equally worth honing. I can attest that I spend much more time debugging than designing, and I'm the best programmer in the world.
It's the digital-vs-analogue issue. It's a lot simpler to model something with discrete blocks when you are able to measure in discrete units, and also easier to repeat, make symmetric, transfer and encode, etc.
But as you say, the price you pay is not being able to shape so minutely or copy precisely, i.e., you have to live with a square wave approximation of the real thing. It's all about tradeoffs. Artsies use analogue media, techies use digital:).
Have you ever seen an employer looking for someone who knows "linked lists" or "use of pointers"?
Certainly. Clueful employers ask about solving problems, and quite frequently a binary tree or a linked list or another basic data structure - which many people still can't grasp - are fundamental in the solution.
They want a prospective employee to have precisely the skills required for the position
Which is why MSCEs and DeVry/ITT churn-outs can still get work, because employers think that's what they need. When employers clue in, that will cease to be the case. But clueful hiring people are, well, if not rare, still seem to be under 50% of interviewers. Big companies - e.g. Microsoft, from what I've heard, and Trilogy by personal experience - know how to hire people with transferable skills and the ability to learn.
This means using this career discontinuity to bone up on the latest-n-greatest in the hot sectors of the industry, to offer a better chance of a finding another great job.
Buzzword compliance won't make you a better programmer, and if you go work for an employer that hires based on such then expect seriously inept coworkers.
If you have a good grounding in CS principles (algorithm efficiency, data structures - linked lists, trees, use of pointers) plus a couple of representative programming languages (C or C++, and perhaps some higher level language like perl or sh) and can go from a problem description to a clean design and readable, documented code - then technologies with new names are unimportant. If you don't, then you'll be playing catch-up all the days of your life and still be woefully bereft of "clue."
Take XML (as in "take my wife - please":). It's just HTML except you get to invent the tags. There are some complications (W3C comes out with a new XML standard daily) but most of them can be ignored. Anyone that's read this paragraph can now put "XML" on their resume, but what's the point? Similarly, once you learn one structured or OOP or functional language, picking up another is easy; learn one RPC system (SUN RPC, CORBA, COM/DCOM,.NET) and the rest are, for the most part, isomorphic.
Here's an article that I wrote a little while ago that might be of assistance.
mtve writes: "Have you ever seen source code that is valid on four languages: Perl, C, Befunge, and BrainF*ck? During the last Perlgolf season famous Perl hacker Jérôme Quelin submittedone such inconceivable masterpiece and now he has published an expanded explanation of his solution. Caution: that text can hurt your mental health. Play Perlgolf!"
Why is it that Slashdot's editors can't fix the mistakes in the above and many, many other articles before posting? Either they don't notice the errors (in which case they should be sacked and replaced), or it's that important to post the article a minute earlier (highly unlikely), or somehow the original wording is considered "sacred" and Not To Be Changed (stupid if true). Come on here. Does the error rate on the front page have to be so high?
The only reason I'm not blaming the submitter (mtve) as well is because it's possible English is not his first language (or even his second). If it is, shame on him too. We all deserve better.
(Oh yes, for those clueless enough to say "What errors", I threw in the required changes in bold. Also, isn't one Perlgolf link enough?)
And if the gap in pay and value between Unix and Windows sys admins is widened, who in their right mind coming out of a CS degree in college (not some fly-by-night certification course) is going to want to use their training to specialize in the market that pays the least?
But wasn't there an interview with the former Project Manager for Excel who said why some of their competitors failed was because they restarted from scratch (i.e netscape).
The manager was Joel Spolsky, the article is here, and his site is Joel on software which has a lot of good articles. Since he's a manager with (it appears) a proven track record, bosses might just listen if programmers go to them with his articles. By the way: the place I work at now gets 0/10 on his software development methods test.
One of my interns at work is a CS undergrad, and I think he's pretty typical of the breed. Talks about Linux all the time to be 'leet, but still gave me a resume done in Word on his pirated Win2K partition.
Real l33t Lunix users use Perl to generate their resume in a variety of different tasty flavours (http://davidrobins.net/Resume.{pl,pdf,txt,html,ps,tex}; Output::RTF module to be finished RSN). I just got sick of updating several files in parallel, so did the "hacker" thing and produced a reusable, extensible solution.
In late breaking news, it was determined that the earth still revolves around the sun. While we have no new developments to report on this subject, we'd just like to remind you that it's still true, just like Copernicus said in 1530.
The Amstrad was my first computer; at around the age of 8 my Dad brought home a CPC 464 (64k, green mono monitor). Told me later he traded it for his old CD player, or somesuch, plus got a sackful of games with it too (cassette, naturally). Eventually got the CPC6128 (128k and a disk compatible with nothing else on earth, and colour!)
I started programming it in BASIC, but that wasn't going anywhere fast (being able to run three times around the house while a sprite is being drawn isn't all that impressive). So I picked up a book on Z80 machine code (mnemonics and encodings, interrupt modes, etc.) and started writing. Hex. No assembler. Mostly RSXes (Resident System Extensions - "|"-commands callable from BASIC, like "|SPRITE,1,x,y"). Teaches you to be very careful about counting your jump offsets. The screen memory was even worse than the VGA's bit planes, though; pixels were interleaved most horribly. (And only a British machine would have a firmware call named KL_TIME_PLEASE....)
I eventually got an assembler, which was heaven compared to hand-hacking the machine code. I was pretty far along on a clone of the early Sierra games (Space Quest etc.). I stopped working on that when we moved from the UK to Canada, and I bought a '386 (writing assembly in DEBUG... but we won't go there, except to say that compared to Z80 the 80x86 instruction set is huge; it even had a multiply instruction - what luxury! No slam on the Z80, though; "small is beautiful").
I also did a lot of reverse engineering of games, especially Sorcery Plus; that taught me a lot (the multimode trick, for those that know the game, is very slick; basically it changes screen modes in the middle of the video refresh).
There are several good CPC emulators, too.
Recently I started on a Windows version of a game I typed in from Amstrad Computer User magazine ("Roland Takes a Running Jump", Z80 assembler, serialized over something like 8 issues; I missed the map creator issue so had to write my own). Cheesy old game, but fun... being able to design levels for it was especially neat. My (non-programmer... I knew no programmers then) friends and I would often design sets of levels for each other to play.
What's my point? Do I have one? If I did it would be: learning to program by typing in machine code in hex was really great (because no other language could ever faze me), and I just wanted to say a few things from experience about a great old Z80 platform.
Slashdot Mirror
czth
PHP or VB/ASP are very simple and easy to learn.
For the love of all that is holy, we don't need more script kiddies or PHP programmer-wannabes. The article's suggestion of Logo is actually pretty good - instant results; for a clean scripting language, give them Ruby (and perhaps they can graduate to Perl later if necessary e.g. for better library support), but dear God let's not create any more Visual Basic users (bah; I refuse to refer to them as programmers). Even Delphi (or Kylix) would be preferable (for RAD Windows development), or Turbo Pascal in DOS (i.e. a Win9x console), and if they can handle Delphi they can handle Visual C++, especially with all the wizards it has. Preferable not to start kids dependent on Windows, though; the suggestion of the web as a first "platform" is a good idea (especially as it can do graphics relatively easily). My kids are gonna use Unix and they're gonna like it.
czth
How about this: write a program that reads an Excel spreadsheet and uploads its content into a database and let me know how long it takes you in PHP.
Simple:
<?phpq l_database_named_foo();
read_excel_spreadsheet_and_upload_contents_to_mys
?>
The next version of PHP might be able to do the same for Postgres and Oracle databases and they are considering the ability to be able to upload the contents to tables named "bar" as well (but you may have to buy the Zend engine for that).
Amen to everything you said (I may add your comment to my page). See also my PHPSucks page. And fortunately my employers do let me use Perl (and C when I think it's justified).
(Hm, since your nick is "Christianfreak" you probably will know what "amen" means, I guess I can leave it in. :)
czth
Documenting the format is a good step, but if Microsoft changes it regularly, Microsoft will retain the advantage that only their programs are compatible with the "standard" set by Microsoft.
Perhaps a requirement for the company "owning" the standard to release changes a fixed amount of time before they release a product would solve that problem. And also it doesn't do them much good to add to the spec if their current software can't read it; they'd break their own software. Or, if their software can read it as "unhandled feature" and still read the document, other readers will be able to do the same. If not, they will be seen to be in violation of the law and be fined or declared invalid as a government supplier, or at least have their new product declared unusable.
czth
Now for the disclaimer. I haven't written either. That's because I'm in the US on a (TN) visa, and I don't believe it would do much good for a non-citizen to write (and it would be dishonest to pretend, even by omission, that I was a citizen). (My fiancée is a US citizen, and I will probably get citizenship eventually, but it will take at least five years after getting a green card.) I could, I suppose, send money to a representative, but (a) I don't believe that's right any more than it is right for corporations to do it and (b) I'd probably get arrested by Big Brother for government subversion by a foreign national ;-).
czth
Remember, there was a time when nobody could imagine the US government waging war against peaceful individuals. Today, the US holds more prisoners per population than any other country in the world -- and the vast majority are drug users or dealers who have initiated no force against anyone.
Yes! Won't somebody think of the poor, peace-loving drug dealers?! OH THE HUMANITY!
czth
Well geez, it wouldn't be a problem in the first place if Python wasn't all fascistic in not allowing:
while line = file.readline(): print line
czth
Nice troll. Very cute. I guess the reason why you haven't been flamed yet is that the Perl programmers can see it's a troll and the PHP users don't care either way.
czth
No, free(x) doesn't assign NULL to x. There's a good reason for this: avoiding unnecessary work. Yes it's minimal work, and lots of systems will define a macro that will do the free and assign NULL too (ISTR MFC has something like that), but it counts over millions of iterations. C++'s delete operator also doesn't assign NULL, for the same reason. Usually the variable you just freed will go out of scope soon anyway, so why waste time writing to it?
If you've worked with assembler you'll see that the often-made jest of C being portable assembler is quite true - working in assembler long enough will have you wishing for exactly the features that C has (functions with local variables and named parameters, the various looping constructs, expressions). And it goes that far and no further, and doesn't hold your hand. For some people and some projects this is fine. For some it isn't, either because the people aren't skilled enough or the tradeoffs in using a higher level language outweigh any possible advantages.
Which language to use is a combination of project applicability, personal preference and the collective experience and skills of you and (if applicable) your team. Saying "C is bad" or "don't use C because people have made mistakes with it" is narrow-minded and stupid. Use what works for you. Unless it's PHP, in which case you need help, quick.
czth
Obviously you've never held a real camera, like one of the Nikon F-series models. Drop the G3, it looks like little bits of plastic would go skittering everwhere. Drop the F5, look out below.
Oh, and by the way, the article wasn't that funny.
czth
PHP is designed more for database applications
No, PHP was designed (and I use that term loosely) to make dynamic web pages, and the database functions were - like most of PHP - thrown in as an afterthought, unlike perl's DBI, which was designed to be OO and have a consistent interface.
although it can be extended (like Perl) to do a lot more. The routines PHP uses for database management have been thoroughly refined and optimized to achieve maximum speed
That's content-free. All languages can be extended, even Visual Basic, and they all (except perhaps INTERCAL) claim that their libraries have been "thoroughly refined and optimized to achieve maximum speed." Of course, some do better than others (see Doug's shootout).
I believe it is generally accepted that for databasing, using PHP will result in faster, more streamlined, and smaller code.
And I believe the earth is flat, but it isn't. Or, BZZZT, WRONG (zero for three)! And databasing? PHP isn't neither particularly fast nor compact.
Sad but true: PHP sucks. It's a half-assed attempt by some lame twits to make a programming language for non-programmers, which is sort of like trying to make a car for non-drivers. Whether they succeeded or failed at that goal, we all still lose, because of the horrible proliferation of all the crap PHP scripts on the world.
Dijkstra would probably revise his famous quote for PHP's benefit to: It is practically impossible to teach good programming style to students that have had prior exposure to PHP; as potential programmers they are mentally mutilated beyond hope of regeneration.
czth
Locals in Perl, are, for almost any practical purpose globals. I really hope you meant my-variables.
Unfortunately for Perl, its my variables are what are generally referred to as "local" (usually stack-based, with a lifetime extending only to the end of the scope in which they were declared[1]) variables, whereas those created (overridden, actually) by local are just temporary scoped replacements for existing globals. Not too long ago someone completing a pre-interview screening task for us (a supposedly experienced perl programmer) managed to (consistently) use local where he should have used my.
[1] Of course their referants may last longer if a reference is kept in an outer scope.
czth
czth
It looks fairly sensible, but the main problem is that (1) good programmers know these things and (2) bad programmers usually can't or won't learn. This makes the audience pretty narrow, i.e., inexperienced programmers with decent raw skills.
And this bit made me laugh (2.5):
What do you do when you start to run out of low-hanging fruit? Well, you can reach higher, or chop the tree down.
It just seems like a funny metaphor (picture it in your head, chopping down the tree is sort of overkill just to get more fruit :), although I understand what he's getting at.
czth
czth
But as you say, the price you pay is not being able to shape so minutely or copy precisely, i.e., you have to live with a square wave approximation of the real thing. It's all about tradeoffs. Artsies use analogue media, techies use digital :).
czth
"There is no new thing new under the sun" (Ecclesiastes).
czth
Have you ever even looked for an IT job?
Several times.
Have you ever seen an employer looking for someone who knows "linked lists" or "use of pointers"?
Certainly. Clueful employers ask about solving problems, and quite frequently a binary tree or a linked list or another basic data structure - which many people still can't grasp - are fundamental in the solution.
They want a prospective employee to have precisely the skills required for the position
Which is why MSCEs and DeVry/ITT churn-outs can still get work, because employers think that's what they need. When employers clue in, that will cease to be the case. But clueful hiring people are, well, if not rare, still seem to be under 50% of interviewers. Big companies - e.g. Microsoft, from what I've heard, and Trilogy by personal experience - know how to hire people with transferable skills and the ability to learn.
czth
Buzzword compliance won't make you a better programmer, and if you go work for an employer that hires based on such then expect seriously inept coworkers.
If you have a good grounding in CS principles (algorithm efficiency, data structures - linked lists, trees, use of pointers) plus a couple of representative programming languages (C or C++, and perhaps some higher level language like perl or sh) and can go from a problem description to a clean design and readable, documented code - then technologies with new names are unimportant. If you don't, then you'll be playing catch-up all the days of your life and still be woefully bereft of "clue."
Take XML (as in "take my wife - please" :). It's just HTML except you get to invent the tags. There are some complications (W3C comes out with a new XML standard daily) but most of them can be ignored. Anyone that's read this paragraph can now put "XML" on their resume, but what's the point? Similarly, once you learn one structured or OOP or functional language, picking up another is easy; learn one RPC system (SUN RPC, CORBA, COM/DCOM, .NET) and the rest are, for the most part, isomorphic.
Here's an article that I wrote a little while ago that might be of assistance.
czth
Why is it that Slashdot's editors can't fix the mistakes in the above and many, many other articles before posting? Either they don't notice the errors (in which case they should be sacked and replaced), or it's that important to post the article a minute earlier (highly unlikely), or somehow the original wording is considered "sacred" and Not To Be Changed (stupid if true). Come on here. Does the error rate on the front page have to be so high?
The only reason I'm not blaming the submitter (mtve) as well is because it's possible English is not his first language (or even his second). If it is, shame on him too. We all deserve better.
(Oh yes, for those clueless enough to say "What errors", I threw in the required changes in bold. Also, isn't one Perlgolf link enough?)
czth
Think of it as evolution in action.
czth
The manager was Joel Spolsky, the article is here, and his site is Joel on software which has a lot of good articles. Since he's a manager with (it appears) a proven track record, bosses might just listen if programmers go to them with his articles. By the way: the place I work at now gets 0/10 on his software development methods test.
czth
Real l33t Lunix users use Perl to generate their resume in a variety of different tasty flavours (http://davidrobins.net/Resume.{pl,pdf,txt,html,ps ,tex}; Output::RTF module to be finished RSN). I just got sick of updating several files in parallel, so did the "hacker" thing and produced a reusable, extensible solution.
czth
In late breaking news, it was determined that the earth still revolves around the sun. While we have no new developments to report on this subject, we'd just like to remind you that it's still true, just like Copernicus said in 1530.
The Amstrad was my first computer; at around the age of 8 my Dad brought home a CPC 464 (64k, green mono monitor). Told me later he traded it for his old CD player, or somesuch, plus got a sackful of games with it too (cassette, naturally). Eventually got the CPC6128 (128k and a disk compatible with nothing else on earth, and colour!)
I started programming it in BASIC, but that wasn't going anywhere fast (being able to run three times around the house while a sprite is being drawn isn't all that impressive). So I picked up a book on Z80 machine code (mnemonics and encodings, interrupt modes, etc.) and started writing. Hex. No assembler. Mostly RSXes (Resident System Extensions - "|"-commands callable from BASIC, like "|SPRITE,1,x,y"). Teaches you to be very careful about counting your jump offsets. The screen memory was even worse than the VGA's bit planes, though; pixels were interleaved most horribly. (And only a British machine would have a firmware call named KL_TIME_PLEASE....)
I eventually got an assembler, which was heaven compared to hand-hacking the machine code. I was pretty far along on a clone of the early Sierra games (Space Quest etc.). I stopped working on that when we moved from the UK to Canada, and I bought a '386 (writing assembly in DEBUG... but we won't go there, except to say that compared to Z80 the 80x86 instruction set is huge; it even had a multiply instruction - what luxury! No slam on the Z80, though; "small is beautiful").
I also did a lot of reverse engineering of games, especially Sorcery Plus; that taught me a lot (the multimode trick, for those that know the game, is very slick; basically it changes screen modes in the middle of the video refresh).
There are several good CPC emulators, too.
Recently I started on a Windows version of a game I typed in from Amstrad Computer User magazine ("Roland Takes a Running Jump", Z80 assembler, serialized over something like 8 issues; I missed the map creator issue so had to write my own). Cheesy old game, but fun... being able to design levels for it was especially neat. My (non-programmer... I knew no programmers then) friends and I would often design sets of levels for each other to play.
What's my point? Do I have one? If I did it would be: learning to program by typing in machine code in hex was really great (because no other language could ever faze me), and I just wanted to say a few things from experience about a great old Z80 platform.
Any other Amstrad hackers out there?