meanwhile, in the anti-matter containment chamber

[sergebroom]

The darn (*) anti-matter containment chamber sprang a new leak.

Believe it or not, something is wrong with the Project from Hell. I am so glad that we spent months thoroughly testing this. My favorite part of the email exchange was the users saying that I had stated that we'd have clean data by January 23. I didn't point out that the way I remember things, they had put a metaphorical gun to my head and had asked if everything would be ready by that date, and all the while it was quite obvious which answer they wanted to hear.

Oh well.
Back into the Jefferies Tube.

----------

(*) 'Darn' is not quite the four-letter word I really have in mind.

Comments

[(Anonymous)]

Why is it that as a casual reader of your blog, after reading that, I thought “Gee, I’m not surprised”?

Because, gee, I’m not surprised.

-scorby

[serge-lj.livejournal.com]

At least, I did find the solution to this situation. I still have to spend days to test it and all that, so as to prove to my boss and our users that I do know what the frak I'm doing, but that doesn't diminish the glow I feel from having figured things out.

[fledgist.livejournal.com]

It does sound like a truly hellish project.

[serge-lj.livejournal.com]

It does make one feel like one is now residing in a rather sulfurous place - ad I'm not talking of the smelly hot-water sources near Mount Lassen in northern California.

I guess I should remind myself to keep things in perspective. There's this mad poetess I know who lives in Edinburgh. She's in charge of demolishing other people's programming... er... I mean, her job is to test things thoroughly. It is my understanding that, as part of her own Infernal Project, she had to figure out how to cram 79 days worth of testing within the month of February - which this year counts 28 days. And she found a way to do it.

Hats off to her.

[evilrooster.livejournal.com]

Hold the parades until I* make it actually happen, Serge. All I've proved is that it's possbile. It's still highly improbable.

*"I", in this case, means "we"; I'll need two project managers and many techies to make this work.

[serge-lj.livejournal.com]

Possible, but highly improbable?

That reminds me of Law & Order when it was still good (1). There was one episode where the defense asks for a conclusion from a forensics person and, not finding said conclusion to his liking, asks if there is another possible explanation for the death of the person his client is accused of killing. The forensics guy's response?

"It's possible he was killed by a death ray from Mars. It's just not very likely."

At least, your group will be given the people and equipment to do the job, from the word go, not at the last minute. That's got to tip the scales in favor of success, more than my Project, where testing was done by yours truly and only one other person, and operating on a server also used by the rest of the team for their own work. (2) Like I said, things are more in favor of your successfully pulling it off, and this without your also finding Tripods right out of your office's window.

My best wishes to you.

----------

(1) Meaning, when Jerry Orbach was still alive.

(2) Our manager's decision to have that server's database upgraded in the middle of all the action is not one that would have him/ber favorably compared to Einstein, or to Feynman.

[evilrooster.livejournal.com]

"At least, your group will be given the people and equipment to do the job, from the word go, not at the last minute."

Sadly, no. This *is* the last minute, the key people are still baulking, and the equipment is shared with another test team who are going to run late, based on past experience.

We may successfully pull it off, but that will mostly involve redefining "success" at the end of the exercise. It's an ancient and honourable tradition in the interaction between software testing and management. Who are we to break it now?

[serge-lj.livejournal.com]

Who are we indeed?

As for defining success, I shall refrain from easy political jokes about missions accomplished. My user at least has the same definition that I do: if there are major discrepancies between what is out there, and what her test scripts say should be out there, then we are not successful. But she is reasonable (and probably just as sick of the frakking Project as I am): a single discrepancy here and there is acceptable. Ten thousands, not so much.

Anyway, software testing makes me think of Asimov's Galactic Empire. They probably stopped upgrading their softwares earlly in Imperial History. How does one set up a testing environment that can simulate millions and millions of worlds? And distributing those upgrades?

"What do you mean? You're still at version 378999.777? Didn't you ever get version 8888999999999999.2?"

Somebody must written a story about that.
Or a sonnet.

[serge-lj.livejournal.com]

Meanwhile, the Project from Hell had one other outstanding issue. I had looked all over the system, trying to think of any exotic chain of events that could have allowed this in spite of all the testing, and every time coming up empty-handed (a bit empty-headed too) after hours of research. This morning, I did some more poking around, and I finally figured it out.

A typo.
A stinking typo.