A Burgundy and Gold Obsession
View RSS Feed

Behind Enemy Lines

Regression Testing

Rate this Entry
Regression Testing is defined in Wikipedia as any type of software testing that seeks to uncover new errors, or regressions, in existing functionality after changes have been made to a system, such as functional enhancements, patches or configuration changes.

It is an integral part of all reputable software development, expected and planned for by the project manager. It has to be as every line of code you write has the potential to cause issues with any other line of code you have already written.

And make no mistakes, bugs do not just happen, they are put there by those creating the program. They may be there from the very first box in the flow chart, get added somewhere in the pseudo-code process or only make an appearance late in process while a build is going through quality assurance testing. Worse, they might only show up when you, the consumer, purchase and install it on your home PC. (No comments about those of you with Macs)

To regression test effectively, QA engineers commonly rerun old test plans, test for specific known bugs that have already been fixed and do general functionality tests to ensure that a program is doing what it is supposed to in a given situation. Any severe bugs that are found must be fixed before the software can go into "release".

The NFL has its own form of regression testing called Game Day.

Coaches and the Front Office are the programmers while players, player acquisition and game plans are the code and the regular season is the test. If you want to get more granular about it, the game each week is a different pass of the full test cycle with the team's final record being its grade and only those teams making the post season make "release".

Our Redskins are failing this test.

And doing so for the second season in a row.

Last year, after an offseason of Front Office and Coaching changing, trades, FA pickups and taking a big LT in the draft, our boys opened the test cycle with a win over Dallas at home and jumped out to a 4-4 record when they hit their midseason Bye week. So far so good, right?

Sure, the team had dropped a couple games we should have won (Houston in Week 2 comes to mind) and played way down to the level of opponents who we should have beaten (the Rams in Week 3) but we were at .500 at the break and feeling pretty good after nice wins over Philly, Green Bay and Chicago.

Some folks were even talking Playoffs

After the bye though, the Skins went into full regression mode, failing nearly every test pass and dropping 6 of 8, including two blowouts to the Eagles and Giants. Certainly injuries played a part as LaRon Landry, the would be defensive MVP of the team and likely shoe in for the Pro Bowl, was lost for the season. So was Clinton Portis on offense. The offensive line was juggled a good bit with the result being that viewers often didnt know who was in or out on any given series.

There were some good things too. Donovan got benched for Rex and, after a rough start, he nearly lead the team to a win in Dallas, hanging 322 yards on the Cowboy D. Ryan Torain laid claim to the starting TB spot after Portis went out, giving us hope for the future there, and Washington played the Giants tough at home in the last game of the year.

But still, you have to say the team regressed during the season.

So in the offseason the coders when back to work, deleting code that didnt function last year, adding new features to the software and generally remaking this v2.0 release of the Mike Shanahan/Bruce Allen Redskins.

Released were Daniels, Rogers, Portis, Rabauch and Dockery. Traded were McNabb and Haynesworth. Drafted was Kerrigan, Jenkins, Hankerson, Helu and about 20 other guys when the Skins started the draft day with only 6 picks. Relatively cheap but serviceable vets not past their prime were brought in like Bowen, Cofield, Atogwa, Wilson, Chester, Gaffney and Hightower.

And then, after a very short training camp, the test pass began anew. Time to see if the v2.0 load was worthy of hope.

And once again it started will with a win in the first game. In fact, this was better win as the Skins were convincing in beating a Giant team reeling from all the pre-season injuries. The Skins had won a game they were supposed to win. Rex looked good, the running game looked good and the defense not only looked good, it looked great as rookie Ryan Kerrigan made so many forget that he wasnt the QB they had wanted to select in the first round by tipping an Eli Manning pass to himself and scoring a TD on that interception.

And yet, since then, it has felt a bit downhill for the Men of Mike. Sure, they pulled out a gutty win in week 2 over the Cardinals but it felt like it was harder than it should have been. Granted, the team had lost games like that before but most felt like they should have handled AZ better than that after the win over NY. This was followed by a typical loss in Dallas where the Skins won in every category but the score, forcing Dallas to kick 6 FGs to win it. Then a win over a bad Rams team that was way too close.

Granted, the argument could have been made at this point that it was the offense holding the team back as the defense wasnt giving much up and was giving team a chance late in games.

Then came Philly and Carolina and our defense got exposed too. The turnovers these two teams had given up to previous opponents hardly materialized and when they did show up against the Eagles, Rex was nice enough to return the favor ASAP.

Sadly, regression testing for the Skins is proving that there are new bugs each week, or maybe they are just bugs we had but didnt know we had. In either case, until the programmers fix those bugs, this version wont be ready for "release".

Updated 10-26-11 at 11:22 PM by Neophyte

Categories
Uncategorized

Comments

Trackbacks

Total Trackbacks 0
Trackback URL: