I've been tasked with testing a Windows program that is essentially a collection of links to various exe files. My job is basically to click all the links to make sure they work (yay!). To makes things a bit easier on myself, I want to create a simple script that will do the clicking for me. I have a programming background (familiar with C/C++, Java), but have never scripted anything before.
Despite googling and searching around on stackoverflow, I wasn't able to find anything that explains what would be involved in creating something like this. I downloaded a program called AutoHotKey (http://www.autohotkey.com/) that seems promising, but before I dive into it I want to see what suggestions you all have.
Thanks!
Never tried autohotkey so I can't give you feedback. But I played around a bit with Sikuli which is agnostic of the UI technology used (working with image recognition). It works fine and you should be able to come up with a working prototype in a couple of hours.
HP Unified Functional Testing (http://www8.hp.com/us/en/software-solutions/unified-functional-testing-automated-testing/) provides similar functionality as Sikuli. As licensed software, it isn't free, but provides a lot more features and support than Sikuli (which is open source) does.
As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance.
Closed 12 years ago.
This may be too opinionated, but what I'm trying to understand why some companies mandate the use of an IDE. In college all I used was vim, although on occasion I used netbeans for use with Java. Netbeans was nice because it did code completion and had some nice templates for configuration of some the stranger services I tried.
Now that my friends are working at big companies, they are telling me that they are required to use eclipse or visual studio, but no one can seem to give a good reason why.
Can someone explain to me why companies force their developers into restricted development environments?
IDE vs Notepad
I've written code in lots of different IDEs and occasionally in notepad. You may totally love notepad, but at some point using notepad is industrial sabatoge, kind of like hiring a gardener who shows up with a spoon instead of a shovel and a thimble instead of a bucket. (But who knows, maybe the most beautiful garden can be made with a spoon and a thimble, but it sure isn't going to be fast)
IDE A vs IDE B
Some IDE's have team and management features. For example, in Visual Studio, there is a screen that finds all the TODO: lines in source code. This allows for a different workflow that may or may not exist in other IDEs. Ditto for source control integration, static code analysis, etc.
IDE old vs IDE new
Big organizations are slow to change. Not really a programming related problem.
Because companies standardize on tools, as well as platforms--if your choice of tools is in conflict with their standards then you can either object, silently use your tool, or use the required tool.
All three are valid; provided your alternative doesn't cause other team-members issues, and provided that you have a valid argument to make (not just whining).
For example: I develop in Visual Studio 2008 as required by work, but use VS2010 whenever possible. Solutions/Projects saved in 2010 can't be opened in 2008 without some manual finagling--so I can't use the tool of my choice because it would cause friction for other developers. We also are required to produce code according to documented standards which are enforced by Resharper and StyleCop--if I switched to a different IDE I would have more difficulty in ensuring the code I produced was up to our standards.
If you're good at using vim and know everything there is to know about it, then there is no reason to switch to an IDE. That said, many IDEs will have lots of useful features that come standard. Maintaining an install of Eclipse is a lot easier than maintaining an install of Vim with plugins X, Y, and Z in order to simulate the same capabilities.
IntelliSense is incredibly useful. I realize that vim has all sorts of auto-completion, but it doesn't give me a list of overloaded methods and argument hints.
Multiple panes to provide class hierarchies/outlines, API reference, console output, etc.. can provide you more information than is available in just multiple text buffers. Yes, I know that you have the quickfix window, but sometimes it's just not enough.
Compile as you type. This doesn't quite work for C++, but is really nice in Java and C#. As soon as I type a line, I'll get feedback on correctness. I'm not arrogant enough as a programmer to assume that I never make syntax errors, or type errors, or forget to have a try/catch, or... (the list goes on)
And the most important of all...
Integrated Debuggers. Double click to set a break point, right click on a variable to set a watch, have a separate pane for changing values on the fly, detailed exception handling all within the same program.
I love vim, and will use it for simple things, or when I want to run a macro, or am stuck with C code. But for more complicated tasks, I'll fire up Eclipse/Visual Studio/Wing.
Sufficiently bad developers are greatly assisted by the adoption of an appropriately-configured IDE. It takes a lot of extra time to help each snowflake through his own custom development environment; if somebody doesn't have the chops to maintain their own dev environment independently, it gets very expensive to support them.
Corporate IT shops are very bad at telling the difference between "sufficiently bad" and "sufficiently good" developers. So they just make everybody do the same thing.
Disclaimer: I use Eclipse and love it.
Theoretically, it would decrease the amount of training needed to get an unexperienced developer to deal with the problems of a particular IDE if all the team uses that one tool.
Anyway, most of the top companies don't force developers to use some specific IDE for now...
I agree with this last way of thinking: You don't need your team to master one particular tool, having team knowledge in many will improve your likelyhood to know better ways to solve a particular roblems.
For me, I use Visual Studio with ReSharper. I cannot be nearly as productive (in .Net) without it. At least, nobody has ever shown me a way to be more productive... Vim, that is great. You can run Vim inside of Visual Studio + R# and get all the niceties that the IDE provides, like code navigation, code completion and refactoring.
Same reason we use a hammer to nail things instead of rocks. It's a better tool.
Now if you are asking why you are forced to use a specific IDE over another, well that's a different topic.
A place that uses .NET will use Visual Studio 99% of the time, at least that's what I've seen. And I haven't found anything out there that is better than Visual Studio for writing .NET applications.
There is much more than code completion into an IDE:
debugging facilities
XML validation
management of servers
automatic imports
syntax checking
graphical modeling
support of popular technologies like Hibernate, TestNG or Spring
integration of source code management
indexing of file names for quick opening
follow "links" in code: implementation, declaration
integration of source code control
searching for classes or methods
code formatting
process monitoring
one click/button debugging/building
method/variable/field/... renaming
etc
Nothing to do with incompetence from the programmers. Anybody would be A LOT less productive using vim for developing a big Java EE application.
How big were you projects at college? A couple of classes in a couple of files? Or rather a couple of hundreds of classes in a couple of hundreds of files?
Today I had the "honor" of looking at a file in a rather large project where the programmer opted to use vi (yes vi, not vim) and a handcrafted commandline compiler call (no make). The file contained on function spanning about 900 lines with a series of if-else-if-else-constructs (because that way you have all your code in one place!!!!!!). Macho-Programmer at his finest.
OK there are very good reasons for enforcing a particular toolset within a production environment:
Companies want to standardize everything so that if an employee leaves they can replace that person with minimal effort.
Commercial IDEs provide a complex enough environment to support a single interface for a variety of development needs and supporting varying levels of code access. For instance the same file-set could be used by the developer, by non-programmers (graphics designers etc.) and document writers.
Combine this with integrated version control and code management without the need of someone learning a particular version control system, all of a sudden IDEs start to look nicer and nicer.
It also streamlines maintenance of build systems in a multi-homed environment.
IDEs are easier to give tutorials to via phone or video, and probably come with those.
etc. etc. and so forth.
The business decision making behind enforcing a standardized environment goes beyond the preference of a single programmer or for that matter perhaps the understanding of the programming team.
Using an IDE helps an employee to work with huge projects with minimal training. Learn a few key combos - and you will comfortably work with multi-thousand-file project in Eclipse, IDE handles most of the work for you under the hood. Just imagine how many years of learning it takes to feel comfortable developing such projects in Vim.
Besides, with an IDE it is easy to support common coding standards across the entire team: just set a couple of options and an IDE will force you to write code in a standardized way.
Plus, IDE gives a few added bonuses like refactoring tools (especially good in Eclipse), integrated debugging (especially good in Visual Studio), intellisense, integrated unit tests, integrated version control system etc.
The advantages and disadvantages of using an IDE also greatly depends on the development platform. Some platforms are geared towards the use of IDEs, others are not. As a rule of thumb, you should use IDE for Java and .Net development (unless you're extremely advanced); you should not use IDE for ruby, python, perl, LISP etc development (unless you're extremely new to these languages and associated frameworks).
Features like these aren't available in vim:
Refactoring
Integrated debuggers
Knowing your code base as an integrated whole (e.g., change a Java class name; have the change reflected in a Spring XML configuration)
Being able to run an app server right inside the IDE so you can deploy and debug your code.
Those are the reasons I choose IntelliJ. I could go back to sticks and bones, but I'd be a lot less productive.
As said before, the question about using an IDE is basicaly productivity. However there is some questions that should be considered by the company when choosing a specific IDE. that includes:
Company culture
Standardize use of tool, making it accessible for all developers. That easies training, reduces costs and improve the speed of learn curve.
Requirements from specific contract. As an example, there are some development packages that are fully supported (i.e. plugins) by some IDE and not by anothers. So, if you are working with the support contract you will want to work with the supported IDE. A concrete example is when you are working with not common OS like VxWorks, where you can work with the Workbench (that truely is an eclipse with lot of specific plugins for eclipse).
Company policy (and also I include the restriction on company budget)
Documentation relating to the IDE
Comunity (A strong one can contribute and develop still further the IDE and help you with your doubts)
Installed Base (no one wants to be the only human to use that IDE on the world)
Support from manufacturers (an IDE about to be discontinued probably will not be a good option)
Requirements from the IDE. (i.e. cross platform or hardware requirements that are incompatible with some machines of the company)
Of course, there is a lot more. However, I think that this short list help you to see that there is some decisions that are not so easy to take, when we are talking about money and some greater companies.
And if you start using your own IDE think what mess will be when another developer start doing maintenance into your code. How do you think will the application be signed at the version manager ? Now think about a company with 30+ developers each using its own IDE (each with its own configuration files, version and all that stuff)...
http://xkcd.com/378/
Real programmers use the best tools available to get the job done. Some companies have licenses for tools but there's nothing saying you can't license/use another IDE and then just have the other IDE open to copy/paste what you've done in your local IDE.
The question is a bit open-ended, perhaps you can make it community wiki...
As you point out, the IDE can be useful, or even a must have, for some operations, like refactoring, or even project exploring: I use Eclipse at my work, on Java projects, and I find very useful to get a list of all occurrences of the usage of a public method or a class in a project. Likely, I appreciate to be able to rename it from where it is defined, and having all these occurrences automatically updated.
The fact I have the JavaDoc displayed when hovering over a name is very nice too. Like autocompletion, jump to a class name, etc.
And, of course, debugging facilities...
Now, usage of Eclipse isn't mandatory in our shop! Some years ago, some people used the Delphi IDE (forgot its name), I tried NetBeans, etc. But I think we de facto standardized on Eclipse, but it was a natural evolution rather than a company policy. And we often just open files in a text editor when we need a quick update...
I was wondering what place scripting has in today's world of IDEs and GUIs.
I'm new to programming and am wondering at what point I should, if at all, open up the PowerShell terminal for a particular task. What do people here use scripting for and how important is it to a modern developer working full time with C++/C#/Java?
In general, you will want to write a script if you ever expect to do the task more than about twice. Making a script has the following advantages:
it's not dependent on human operation (humans are fallible)
it's repeatable
scripts can be stored in source control
This is well suited to tasks such as build, deployment, and automated testing.
IDEs and GUIs don't handle the testing or deployment parts of the development process well quite yet.
Scripting is also great when you want to do something quickly just to see. Personally, I use my Python interpreter as a super-calculator all the time. I also use it to parse files, combined with regular expressions.
Scripting is also often used to allow clients to easily extend a system's behavior. Using a script language, a compiler (or even an IDE) is not necessary.
See Extending Packages with Scripting for an example in SQL Server 2008.
Another point to add is that several servers which are highly used do not have any GUI. Everything is run from a terminal. Almost everything is run through a script. While GUIs are nice, they are not going to be used all the time.
Which is better for quickly developing small utilities? AutoIt or AutoHotKey or something else?
I need to develop some small software for which I think using some professional suite like Visual Studio will be overkill. Most of the macro recording tools like AutoIt or AutoHotKey provide enough power to write decent application. Plus they are small and free.
Which option will be good? Using one of these tools or using some other small/free compiler?
They are both pretty good, and can do most of the same things... but I typically use AutoIt over AutoHotKey for the following reasons:
If you need a GUI at all, then you'll need to use AutoIt.
I prefer the syntax of AutoIt over AutoHotKey's. IMO, it is much easier to follow, and pick back up when you need to go back into your code later.
I like that AutoIt makes it very easy to change version numbers, icons, and compression of your compiled scripts (when compiling to exe). The default compression will sometimes cause false positives with some anti-virus software.
Ultimately, though, it really depends on what your requirements are for your small utilities.
Both are really good pieces of software and you will only decide by trying both of them out with tests that check if you can create your small bits of software.
Me personally I would probably investigate AutoIt in depth and if it started to mess me about I would then check AutoHotKey and others.
You can pretty much do whatever you can in one in the other, but I think it breaks down to: (1) short, keyboard shortcut driven utils and (2) more complex, UI driven scripts. For example:
I use AutoHotKey scripts every few minutes, and use AutoHotKey here over AutoIT because it's snappier to launch (ie via keyboard shortcuts).
I use AutoIT scripts every once in a while to do more complex things, such as to run multi-stage tasks that require a UI... or for when I would like to easily produce a compiled .exe
I realize you can script Microsoft Office apps, but I'm looking for something more general that I can apply to other apps, such as Adobe Acrobat, web browsers and other apps with no scripting ability.
I've used AutoIt but it's a bit clunky, especially when trying to debug why a script fails or stalls. Also, getting the timing of actions, such as clicking a button, or selecting a menu item correctly can be a pain.
Are there build tools that could be used for this purpose?
I recommend AutoHotKey. Its syntax is not pretty, but most of the times you don't have to concoct your own scripts, because its community is so large and well organized. Even if you do, the documentation is extensive and good and the forums will answer your questions quickly. The developer is active and responsive, which means that bugs get fixed quickly, and new features are being considered and added.
Since I began using AHK I don't imagine doing without it - it allows to make life on Windows simpler in so many ways.
You can also employ the COM interface from Python and other scripting languages. It is more complex, but you get to use a more powerful language.
I love AutoHotkey (small k...) too, but beside its odd syntax, it has the same lack of debugging tools...
Basically, that's "show msgbox alerts, send strings to file or debugview, trace". Which is OK for most cases, since you rarely write long and complex applications with these tools.
In both tools, and probably all macro softwares, "timing of actions" will be hard to get anyway, because events are asynchronous: most of the time, you don't wait a given time, but you wait for a window to appear. Hoping it is the right one!
There are other automation tools, like Ranorex (I didn't tested it), you can even use some scripting language (Lua, Python) with a library to send messages (WM_XXX) and another to call WinAPI... But tools like AutoIt and AutoHotkey have the advantage of having been extensively tested, so they can handle a large number of behaviors/issues (like waiting for clipboard data to be available, etc.).
Might be a bit much for your needs but AutoMate is very robust and easy to use. Doesn't require any scripting skills as most tasks can be constructed via drag and drop http://www.networkautomation.com/sales/scripting/