Re: minimanual.pdf -- comments requested
- Posted by petelomax Oct 14, 2015
- 2240 views
That is better, but I still don't like "oE", even in plain text. I can only assume it is pronounced like a child being sick! Why not just use Euphoria? On a fresh install, can you not just double-click on wee.exw? The opening example is five times too complicated. Start with message_box("This is Euphoria","My first program",MB_OK), or won't that work on Linux? Maybe puts(1,"Welcome to programming in Euphoria") any_key(), or string name = prompt_string("Enter your name") puts(1,"Hello "&name) any_key().
I don't like "converted into a binary application". I would say something like "of course you can share the source code of your application, but then it can only be run where openEuphoria has already been installed. Use Run|Bind from the Wee menu to create a standard/standalone executable application (intro on Linux, intro.exe on Windows) that anyone can run. I'd also drop the "A compiler makes a better performing application" from the Flexible section: it screams that interpretation is horribly slow, and it's not usually the real reason for compiling (distribution is).
oE is powerful. I get it, but a) it is just bragging and b) it is not really communicating anything. Perhaps a better approach is to say something like some languages are deliberately designed to be difficult, some have so many ways of doing things they inadvertently become difficult. Euphoria keeps things simple so you can concentrate on the problem you are trying to solve, rather than add a whole new layer of problems as part of the solution. There are many freely available examples that prove keeping things simple does not have to sacrifice being powerful and flexible.
It no longer works to promote Eu as "not OOP". Rather Eu provides many of the benefits of OOP but in a much simpler way. For instance it is easy to place some data and all the routines that can operate on it in a separate file, and limit what routines (or data) can be invoked (or seen) from the rest of the application. You can easily attach any routine_id to any data; there is no need to declare a class or carefully construct a complicated hierarchy of inheritance just to get the right routines to be associated with some particular data.
The lack of an index/table of contents limits the usefulness of this severely, imo.
page 7 "que". "9/5* C +32" (wierd spacing)
Page 13: spurious !
Page 16: "& str[6]" should probably be "& str[6..6]" or better yet "& str[6..$]"
Page 22: end if/print wrong indent *2
You can get your own back by reviewing my effort, if you can handle chm format files: https://bitbucket.org/petelomax/phix then navigate source -> docs -> phix -> phix.chm
Pete