Saturday, August 26, 2006
Installation of InnerLoop
One of the things I like best about InnerLoop is how fast and easy it can be installed.
The download is pretty big, but not as large as installing Specman or NCsim. I can download a new version of InnerLoop and install it in less than 7 minutes. Because InnerLoop doesn't depend on the version of Linux or the version of Java it's never any problem regardless of the type of machine I try to use.
Just extract the tar file and run the script named install.
When the legal verbage comes out as a more just hit q and type yes and enter the directory to install and it's done, that's it.
No configurator to run and no more funny questions.
I would guess Windows installation is different but I have never used InnerLoop on windows, hopefully it's just as easy.
There's something to be said for speed and simplicity when it comes to InnerLoop.
One of the things I like best about InnerLoop is how fast and easy it can be installed.
The download is pretty big, but not as large as installing Specman or NCsim. I can download a new version of InnerLoop and install it in less than 7 minutes. Because InnerLoop doesn't depend on the version of Linux or the version of Java it's never any problem regardless of the type of machine I try to use.
Just extract the tar file and run the script named install.
When the legal verbage comes out as a more just hit q and type yes and enter the directory to install and it's done, that's it.
No configurator to run and no more funny questions.
I would guess Windows installation is different but I have never used InnerLoop on windows, hopefully it's just as easy.
There's something to be said for speed and simplicity when it comes to InnerLoop.
Wednesday, August 16, 2006
Screenshots
Somebody asked me why I don't publish screenshots of all the cool features I describe in this blog.
There are two reasons:
1. It takes too long and I don't have time.
2. If I show the details you won't have to do it yourself to see it. The fun is in learning, trying things and seeing what happens.
Somebody asked me why I don't publish screenshots of all the cool features I describe in this blog.
There are two reasons:
1. It takes too long and I don't have time.
2. If I show the details you won't have to do it yourself to see it. The fun is in learning, trying things and seeing what happens.
Coding Without Errors
After entering a bunch of new code it's usually good to see if it will compile to get rid of all the syntax errors. Last time I discussed using statement templates to avoid typos in the code. This process of edit, compile, edit, compile is part of the reason InnerLoop got it's name. Engineers spend all of their time traveling the inner loop of edit, compile, run, and debug (that is all of their time they are not sending e-mail, surfing the web to read the news, or sitting in meetings).
Another way to cut down the number of iterations is to avoid entering any syntax errors. Of course, it's not so easy. Sure, you could spend more time to inspect the code and hope you can see more of the errors, but that just ends up taking more time.
InnerLoop offers an "Errors" tab that dynamically parses the e code as it is entered. If the parser detects any syntax errors they will show up in the Errors tab. Before even trying to compile anything it's very easy to check the Errors tab and have a look for syntax errors. The parser can't catch every error (yet), but over time I bet it will catch most of them.
Give it a try. See if you can enter new code that is syntax error free (without staring at it for hours) by looking at the Errors tab. It's pretty cool, and I bet vim can't do that.
After entering a bunch of new code it's usually good to see if it will compile to get rid of all the syntax errors. Last time I discussed using statement templates to avoid typos in the code. This process of edit, compile, edit, compile is part of the reason InnerLoop got it's name. Engineers spend all of their time traveling the inner loop of edit, compile, run, and debug (that is all of their time they are not sending e-mail, surfing the web to read the news, or sitting in meetings).
Another way to cut down the number of iterations is to avoid entering any syntax errors. Of course, it's not so easy. Sure, you could spend more time to inspect the code and hope you can see more of the errors, but that just ends up taking more time.
InnerLoop offers an "Errors" tab that dynamically parses the e code as it is entered. If the parser detects any syntax errors they will show up in the Errors tab. Before even trying to compile anything it's very easy to check the Errors tab and have a look for syntax errors. The parser can't catch every error (yet), but over time I bet it will catch most of them.
Give it a try. See if you can enter new code that is syntax error free (without staring at it for hours) by looking at the Errors tab. It's pretty cool, and I bet vim can't do that.
Saturday, August 12, 2006
Statement Templates
I have been working with e for some time, but since my background is much heavier in C and Verilog I still have brain blocks when coding in 3-4 languages every day. At such times I turn to the Statement menu in InnerLoop. The statements menu has templates for all of the common e statements with very easy entry boxes to type in the required names to fill in the template and lay in down in the code. This saves a lot of thinking and looking for similar code to do cut-and-paste when I can't remember the exact syntax of e. It also makes sure I don't end up with syntax errors when I go to compile.
Do you ever try to compile and get a syntax error? Unfortunately, I still do also, but I'm on the way to completely eliminating all syntax errors before compiling with InnerLoop. Next time I'll show you how.
Make sure to check out the Statements menu and request support for additional entries you think are useful.
I have been working with e for some time, but since my background is much heavier in C and Verilog I still have brain blocks when coding in 3-4 languages every day. At such times I turn to the Statement menu in InnerLoop. The statements menu has templates for all of the common e statements with very easy entry boxes to type in the required names to fill in the template and lay in down in the code. This saves a lot of thinking and looking for similar code to do cut-and-paste when I can't remember the exact syntax of e. It also makes sure I don't end up with syntax errors when I go to compile.
Do you ever try to compile and get a syntax error? Unfortunately, I still do also, but I'm on the way to completely eliminating all syntax errors before compiling with InnerLoop. Next time I'll show you how.
Make sure to check out the Statements menu and request support for additional entries you think are useful.
Sunday, August 06, 2006
InnerLoop as a Shell
One of the cool things about InnerLoop is you can do almost everything without ever leaving. To do most anything in Linux requires some typing at the shell prompt. InnerLoop has a very easy way to do it. Highlight the Build tab on the bottom of the tool. Initially, there is nothing there and if you don't use the Build options it may not have occurred to you that this is a shell. To try it out type any command in the box, for example pwd. You will get a nice shell prompt that you can use without having to leave InnerLoop and search for an xterm.
For me, I don't normally use the build system that is provided by InnerLoop. Unfortunately, not everybody I work with is enlightened enough to use InnerLoop and scripts and Makefiles tend to be the way things go. However, this doesn't stop me from using the Build window as a shell. When I want to check my e code and generate a stubs file to make sure life is good, I go to the Build window and type the script or make target to do this. If I have any errors (a rare situation) I can click on them and go directly to the source code and fix them.
I notice some users of InnerLoop have a "never leave" philosophy and other tend to leave when it's time to compile and run. The beauty is that it works either way and any way in between. I tend to spent lots of time in InnerLoop and stay in all the way until I'm ready to simulate and then I leave and go for specify or an xterm to run simulation.
Give the Build window a try as a shell today and see you how you like it.
One of the cool things about InnerLoop is you can do almost everything without ever leaving. To do most anything in Linux requires some typing at the shell prompt. InnerLoop has a very easy way to do it. Highlight the Build tab on the bottom of the tool. Initially, there is nothing there and if you don't use the Build options it may not have occurred to you that this is a shell. To try it out type any command in the box, for example pwd. You will get a nice shell prompt that you can use without having to leave InnerLoop and search for an xterm.
For me, I don't normally use the build system that is provided by InnerLoop. Unfortunately, not everybody I work with is enlightened enough to use InnerLoop and scripts and Makefiles tend to be the way things go. However, this doesn't stop me from using the Build window as a shell. When I want to check my e code and generate a stubs file to make sure life is good, I go to the Build window and type the script or make target to do this. If I have any errors (a rare situation) I can click on them and go directly to the source code and fix them.
I notice some users of InnerLoop have a "never leave" philosophy and other tend to leave when it's time to compile and run. The beauty is that it works either way and any way in between. I tend to spent lots of time in InnerLoop and stay in all the way until I'm ready to simulate and then I leave and go for specify or an xterm to run simulation.
Give the Build window a try as a shell today and see you how you like it.
