Talk:Test a function: Difference between revisions
Content added Content deleted
(I think test frameworks are a good idea.) |
m (fix unsigned comment) |
||
| Line 4: | Line 4: | ||
:I'm tempted to disagree with such a change. Having a testing framework makes it easier to test the language implementation itself; they're often among the biggest consumers of such frameworks. So, implementing a framework (which doesn't actually need to be ''that'' complex) that lets you satisfy this task means that you can have a higher-quality language implementation. View it as a hint for how to make the jump to the next level of software engineering quality. (For example, [[Tcl]]'s test suite exercises virtually the entire language implementation, including a very large fraction of the failure paths, and it makes it much easier to know early that maintenance hasn't screwed things up.) |
:I'm tempted to disagree with such a change. Having a testing framework makes it easier to test the language implementation itself; they're often among the biggest consumers of such frameworks. So, implementing a framework (which doesn't actually need to be ''that'' complex) that lets you satisfy this task means that you can have a higher-quality language implementation. View it as a hint for how to make the jump to the next level of software engineering quality. (For example, [[Tcl]]'s test suite exercises virtually the entire language implementation, including a very large fraction of the failure paths, and it makes it much easier to know early that maintenance hasn't screwed things up.) |
||
:On the other hand, a test framework really doesn't need to be more than a few functions; one to run tests, one to compare results for equality, one to print a summary at the end. |
:On the other hand, a test framework really doesn't need to be more than a few functions; one to run tests, one to compare results for equality, one to print a summary at the end. —[[User:Dkf|Dkf]] 23:36, 15 June 2009 (UTC) |
||
Revision as of 01:07, 16 June 2009
How about just write a test for the palindrome task, if your language doesn't have large testing frameworks?
Shouldn't tasks lead toward implementation rather than omitting? --Tinku99 22:52, 15 June 2009 (UTC)
- I'm tempted to disagree with such a change. Having a testing framework makes it easier to test the language implementation itself; they're often among the biggest consumers of such frameworks. So, implementing a framework (which doesn't actually need to be that complex) that lets you satisfy this task means that you can have a higher-quality language implementation. View it as a hint for how to make the jump to the next level of software engineering quality. (For example, Tcl's test suite exercises virtually the entire language implementation, including a very large fraction of the failure paths, and it makes it much easier to know early that maintenance hasn't screwed things up.)
- On the other hand, a test framework really doesn't need to be more than a few functions; one to run tests, one to compare results for equality, one to print a summary at the end. —Dkf 23:36, 15 June 2009 (UTC)