The Daily WTF: Curious Perversions in Information Technology

2006-06-28 Reply Admin

Anonymous:

Harald:
So, do what we did: simply hide the version/build numbers on the dialog, and when the user presses a certain keyboard combination, reveal the info.
So, it's something like: Support: Ok, now we have to get your version number. Press up, down, up, down, left, right, left, right, a, b, a, b, l, r, start, start, select, select?

heard over phone: Finish Him!!!

Customer: Ok I've done that. What do I do now? Its just frozen Word with what looked like an ice blast, then punched it into 100 pieces.

*cue 90's hardcore techno music*

MORTAL KODEBAAAASE!!

(TEST... YOUR BUILD!)

Etc, etc. joke.Run(Into_Ground);

--
$0.02

2006-06-28 Reply Admin

Yes !!!! Finally a subject that the managers can talk about : Version number...it shows all the information you need ...from it you can deduct how old the application is, and how stable it is...

Only one problem: If you remove that info...what wil they talk about next?

If I were you, I wouldn't stop there, I would also remove any tags from CVS (or whetever you use as Source Control repository), and then screw around with the build scripts and test scripts...and then quit that f*cking job....

After a while maybe they would call you for assistance..and than ask them..."Ok, can you tell me What version number of the application was?"

2006-06-28 Reply Admin

Reading this, I was reminded of that old joke where $supreme_being was a software developer. Okay, you have this runtime library called "DNA" that you use to develop a new localized version of Human™. You get nine months to code the software, and one attempt to compile one build. It cannot be restarted after a crash, needs a very robust anti-virus subsystem, and takes up to 20 years to allocate memory for it.

I guess now that we have cloning, we'll soon have versions of people. Will the version/build info be, as posed by a certain Monty Python sketch I remember, "tattooed on the back of the neck"?

--
$0.02

DZ-Jay · 2006-06-28 Reply Admin

From: A----- Selvan
Sent: Wednesday, March 29, 2006 10:10 AM
To: {IT Director}
Cc: {Company CEO}
Subject: Fwd: Software Version

Please revoke any e-mail privileges from {Marketing Director} and
{Product Manager} as they have no real use for it, and seem to be
abusing the system with trivial, non-sensical application change
requests.  It is really affecting the productivity, not to mention
the morale, of the development and support groups.

Thank you,
A-----

2006-06-28 Reply Admin

Phil the ruler of heck:
Why not use concurrent version-numbering schemes at the same time, the way Sun Microsystems does it? SunOS 5.6 = Solaris 2.6, SunOS 5.7 = Solaris 7 (not 2.7, but there was no Solaris 3 through 6 either), Java 1.3 and Java 1.4 = Java 2, Java 1.5 = Java 5.0, and on top of all that, Enterprisey Java 1.4 now comes with Java 5.0 if I read their Web page right...

That's most concise explanation of Sun's versioning that I've ever seen. A career in marketing awaits!

CAPTCHA: perfection

Sinedia · 2006-06-28 Reply Admin

Bus Raker:
If s/he had suggested something like 'let's use the release date to indentify the product instead' it might make some sense. I mean, who really has two releases in the same day?

We do... quite regularly actually :(

2006-06-28 Reply Admin

A good response from A. may have been:

"What if I change 2.8 (build 448) to 2800448? I mean, look how well Microsoft has done with really big version numbers like 2000, 2003, 2005, etc. We can do them one better and go to 7 digits!"

2006-06-28 Reply Admin

Anonymous:
Or one could just do like the TV industry and use roman numerals... I wasn't until the year 2000 roll over that most people figured out what all the crap at the end of the show was. In a few years they'll all forget and their will be enough digits that it will fall back into a senseless blur of Cs and Vs again.

Actually it's 84 (LXXXIV) years till we get a C in there again, hell it's 34 (XXXIV) years till we get an L.

2006-06-28 Reply Admin

Start at # 666 and tell the first caller that the new system was a real devil to put in.

2006-06-28 Reply Admin

Anonymous:
Phil the ruler of heck:
Why not use concurrent version-numbering schemes at the same time, the way Sun Microsystems does it? SunOS 5.6 = Solaris 2.6, SunOS 5.7 = Solaris 7 (not 2.7, but there was no Solaris 3 through 6 either), Java 1.3 and Java 1.4 = Java 2, Java 1.5 = Java 5.0, and on top of all that, Enterprisey Java 1.4 now comes with Java 5.0 if I read their Web page right...

That's most concise explanation of Sun's versioning that I've ever seen. A career in marketing awaits!

Hey, I worked for a company that made PBX and firewall reporting software. We went from version 4.5 to version 1 because we added a lot a features and changed the name of the product. Still the same code, but a new name a new features, and product management didn't want to start at version 5 with the new name since there hadn't been a version 1,2,3 or 4.

However, going from 4.2 to 4.5 was just fine.

A version number is only as good as the marketing department wants it to be.

2006-06-28 Reply Admin

Phil the ruler of heck:
Antti (build 06-28-2006 8:55 AM):

Just how many releases it must have taken to create Oracle 9.4.0.1.0? Nearly 100000! That's one release every day since the middle of the 18th century.

As far as I know, there is no Oracle 9.4 but there may be a 9.1.0.4.0, did you perhaps mean that one?

Wow! Larry Ellison is older than I thought!

:)

CAPTCHA: Pizza. Great, now I'm hungry.

2006-06-28 Reply Admin

Anonymous:
Well from a marketing perspective there is a point to what these guys are saying.

No, there isn't. I defy you to find me a single user who ever came to this conclusion. Come to think of it, I defy you to find more than a handful of users who even know what/where the version number is usually stored, and who even notice its existence.

This is an example of a prize idiot assuming he's of average intelligence/experience. "I (just about) know this isn't the case", goes the thinking, "but I'm quite bright - if I turn down my smarts three notches, I might think... this. OMG! Change it immediately!!!!!!".

The flaw in this reasoning is that the two people concerned are actually well towards the bottom of the pile in terms of smarts, and *even* *they* know not to worry about the version number.

The chances of finding someone else who eschews software "because it took them a few goes to get right" is therefore negligible. And TBH, even if the said user-bovine does turn up, how's he going to run the software anyway on his right-first-time Windows 1.0 box?

Anonymous:
I don't consider this one a genuine WTF. This is just marketing and tech miscommunicating, nothing really stupid, both sides have a point here.

Not even close. Marketing are working from a flawed and baseless position of "what if" they found a user stupid enough to object to version numbers, but only in the case of their product (or they'd still be using Windows 1.0, IE/Mozilla 1.0, etc, etc, etc). There is no actual problem here, but they're still wasting developer time making them jump through trivial hoops so it looks like they're involved, on-the-ball and Taking Charge.

This is a classic case of someone with no understanding of the issue, no appreciation of the fact that everyone already knows better than them, no desire to listen to reason from people who do know the subject *and* the power to effect change in the product... and that's a stupiud and dangerous combination.

I worked for acompany like this for nigh-on two years, and the problem is that the more you give in to these occurrances of idiocy (or at least treat them seriously), the more they, well, occur.

ParkinT · 2006-06-28 Reply Admin

My office-mate caught me saying, "WHAT THE F*CK!?" out loud!!!

2006-06-28 Reply Admin

This is a ridiculous request in the first place, and it makes sense that it would come from someone in marketing. The WTF I see is that the project manage actually conceeds to this point. If customers are looking for a product that's reliable amd they actually understand how builds/versions are related, then they would know that the higher the number the more updated and quality tested a product it is. If I were in their marketing department I would want the version/build on the splash screen of the product for everyone to see. That's how our product is marketed anyway.

2006-06-28 Reply Admin

Anonymous:
Actually it's 84 (LXXXIV) years till we get a C in there again, hell it's 34 (XXXIV) years till we get an L.

However "XL" would be a nicely marketable version number, of course having to be followed by XXL, XXXL and "SuperSize" versions.

2006-06-28 Reply Admin

> No, there isn't. I defy you to find me a single user who ever came to this conclusion.

I've worked with people who said "IE 5 is better than Netscape 4 because 5 > 4".

When you're trying to sell software into a medium-to-large organization, you often need buy-in from several people, as well as at least compliance from several others. If any of them state something boldly like "5 > 4," that can paralyze the rest of the decision-makers. Is standing up to the loudmouth really worth it? Hell, it's just a job.

And before you think this is just BigCo's problem, if I'm trying to sell to them it becomes my problem too.

2006-06-28 Reply Admin

So just call it "Dapper Drake".

2006-06-28 Reply Admin

Umm...as another (and probably another and another) person mentioned, it's easy.

Just reformat the string.

Instead of "build 448", make it

v 2.8.3.448

I did something like that in a previous life. First two numbers were major and minor release versions. Third number was an internal code for platform (1=windows, 2=mac, 3=linux or somesuch), and the fourth # was the build #.

Simple.

2006-06-28 Reply Admin

Good job these people weren't the marketing team for 7-Up.

Cheers,

Rich.

Phil the ruler of heck · 2006-06-28 Reply Admin

Phil the ruler of heck (build 06-28-2006 11:45 AM):
Antti (build 06-28-2006 8:55 AM):
Just how many releases it must have taken to create Oracle 9.4.0.1.0? Nearly 100000! That's one release every day since the middle of the 18th century.

As far as I know, there is no Oracle 9.4 but there may be a 9.1.0.4.0, did you perhaps mean that one?

Oops, my mistake, there is no Oracle 9.1 either. Oracle 9 Release 1 is 9.0 and Oracle 9 Release 2 is 9.2, see http://en.wikipedia.org/wiki/Oracle_database#Version_numbering_conventions. (They're almost, but not quite, as good as Sun's version numbers.)

So you have the choice of 9.0.1.4.0 or 9.2.0.4.0, which will it be?

2006-06-28 Reply Admin

Anonymous:
I raise holy hell if a COTS vendor doesn't furnish adequate change documentation with a new version, and frankly, build dates are inadequately descriptive for most large pieces of software, because there is so much useful information you can get into a version string and so little that you can put in a build date. A build date gives you a point in time; the version string actually tells you something.

29Jun2006 - added library function getWTF() 4.11b2601 - added library function getWTF()

WOW, look at all the extra information contained in the version string as opposed to the date string. Maybe the customer will then sell that incredibly useful information to the competetion and drive us out of business!

Or do you hash the changelog and use that as your version string? 4.11b2006BkdFDhbdFikodfklTeDiDbbmsOOerhDFNHKoFihsidDIIFDHsdoSDFIJsddsfhOUIFDHOIsdFIhDOIfDFiodfpioFDZOPIFADHSUJDFHEBLKCJNELDFKJEHFLEHOAISDFHOW

2006-06-28 Reply Admin

So my XP version is Build 2600.xpsp_sp2_gdr.050301-1519

I checked with Microsoft, they did change XP 260,020,503,011,519 times.

2006-06-28 Reply Admin

It seems like the real problem marketing is complaining about is that the customer can figure out what the version number means. The easiest solution would be to separate the component numbers with dashes and call it the "support code" in the about box instead of "version number".

2006-06-28 Reply Admin

Oh this could be happening in my company aswell :(
2 years ago I had to some changes to existing software...
The version I started working on was 1.1.
After I was done I incremented it to 1.2 (It were minor changes)
Then marketing made me change it to 1.11... because it looked better!
I fought it for a while and finally gave up... it has been released as 1.11.
Now customers ask whatever happened to 1.2, 1.3, 1.4... and why they did not got those versions even if they got a meantenance contract :)

John Bigboote · 2006-06-28 Reply Admin

Anonymous:
This is a ridiculous request in the first place, and it makes sense that it would come from someone in marketing. The WTF I see is that the project manage actually conceeds to this point.

Most of the project managers I've worked with are just cat-herders who are interested in two things:

Delivering the product in some semblance of a reasonable timeline
Getting people to shut up and stop arguing

zamies · 2006-06-28 Reply Admin

Alex Papadimoulis:

As a software development manager, A. Selvan sees his fair share of WTF's. Though most are suited for the likes of Dilbert comic strips, every now and then he'll come across a story that best fits here. Today I'd like to share one of these stories in the form of an email conversation between A. and the new Product Manager...

______________________________________________________________________
From: {Product Manager}
Sent: Tuesday, March 28, 2006 8:41 AM
To: A----- Selvan
Cc: {Marketing Director}
Subject: Software Version

A-----, 
I've noticed that in several places (most prominently, Help-About), 
there is the product version, build number, etc. Please have your team 
remove this information in our next release.

______________________________________________________________________
From: A----- Selvan
Sent: Tuesday, March 28, 2006 9:03 AM
To: {Product Manager}
Cc: {Marketing Director}
Subject: Re: Software Version

The product version and build number are necessary to identify what
release the customer has. This is also key in our testing and quality
process.

______________________________________________________________________
From: {Product Manager}
Sent: Tuesday, March 28, 2006 1:29 PM
To: A----- Selvan
Cc: {Marketing Director}
Subject: Re: Software Version

Couldn't you determine this information some other way? We don't want 
the customers knowing this information and need it removed.

______________________________________________________________________
From: A----- Selvan
Sent: Tuesday, March 28, 2006 1:45 PM
To: {Product Manager}
Cc: {Marketing Director}
Subject: Re: Software Version

We really can't remove it from the product, but, if you'd like, we can
hide it further. However, it is very discrete as is: it only appears in 
the About dialog, release notes, and change log. But no matter what, we 
will need to keep it accessible for support. 

Either way, I don't really understand the request. Almost every software 
program ever made has release information like this. Why would you 
want it removed?

______________________________________________________________________
From: {Product Manager}
Sent: Wednesday, March 29, 2006 10:04 AM
To: A----- Selvan
Cc: {Marketing Director}
Subject: Re: Software Version

We can't be at all accessible to the client, we'll just have to find
another way. 

The reason is simple: when they see  "Version 2.8 (build 448)," they 
will think that it took us 28 releases and over *four hundred* builds 
to get right. 

______________________________________________________________________
From: {Marketing Director}
Sent: Wednesday, March 29, 2006 10:08 AM
To: A----- Selvan, {Product Manager}
Subject: Re: Software Version

This is a good point. This is *not* the information we should be 
conveying to our customers. A-----, please make sure it's removed in
the next release.

Hilarious... although?
In a company I once worked for a manager had the same kind of bowel movement.
We used to manage changes with a protocol called Request For Change --> the manager insisted it would be Request For Enhancement.

I think manager are most likely to eat their own excrements if you call it dark pudding instead of SHIT

2006-06-28 Reply Admin

This is not a WTF. We have had customers who looked at the version number of our product and complain about it. We solved the problem by cutting off the build number when we compile for general release.

2006-06-28 Reply Admin

Well, you can version it as 2006.06.17.23 than, every one will be happy.

2006-06-28 Reply Admin

Ah, Clipper. I'd almost managed to suppress that. I had a program that killed the compiler, no message, no line number, no progress information, no nothin. Had to get a prior working version from source control (diskette in my other jeans) and apply changes one at a time until it croaked again. Problem was a missing comma. But the "real WTF" on that project was the management. A story for another day. I'll just mentioned the two culprits got promotions out of the thing, I theorize just so they could never make THAT particular mistake again.

2006-06-28 Reply Admin

I read in Wikipedia that the Oracle RDMS was first release as version 2, so it would sound important...

2006-06-28 Reply Admin

Anonymous:
Silly as the request is one could rename 'Version' to 'Release Code' and then take some sort of hash of the version and build number.

It will be a pain for support to have a little de-hasher program and a pain for the customers to recite some long string of senseless letters and numbers but hey we're all used to that from trying to register WinXP anyway

This is not a WTF. Managers and marketers aren't always idiots, but I think programmers are too quick to assume that they're always right. Sure, you need release information for support purposes, but its not inconceivable that showing end users the number of builds between releases is undesirable.

Of course, a hash of the build number WOULD be a WTF, because there's no such thing as "dehash" -- by definition. Of course, some other kind of encoding would make perfect sense. How about hex? Or some kind of alphanumeric cypher?

Plain: Release 2.8 (Build 448)
Hex: Version Code 2-8-1C0
Alpha substitution: Version Code B-H-DDH

Of course, this probably wouldn't occur to a product manager or marketing manager because they're not computer scientists. The real WTF is that the developer didn't come up with an obvious solution to a trivial problem and helpfully suggest it as an alternative to the manager.

Nothing will ever beat the PHP WTF from a couple days ago. Everything from now on is just denouement.

2006-06-28 Reply Admin

Anonymous:
If customers are looking for a product that's reliable amd they actually understand how builds/versions are related, then they would know that the higher the number the more updated and quality tested a product it is.

This is absolutely untrue, in 2 different ways. First of all, build number has little correlation to quality or testing. There's no way to know, strictly from a build number, either of those things. Different software projects have totally different build strategies. Some places build daily (overnight even) and others coordinate major builds just once every few months.

Secondly, even if your supposition about correlation between build number and quality or testing was true, there's still no reason to ever assume that the end user knows anything about anything, let anything about software development life cycle. In fact, its a very dangerous assumption.

2006-06-28 Reply Admin

>>...amd they actually understand how builds/versions are related...

I mispelled 'and', but I would assume you knew what I meant.

I certainly wouldn't assume that they had any understanding of it, but management in this case is implying that their customer base might have some knowledge of this. If that's the case then they wouldn't be deterred by higher numbers, while it may not have anything to do with testing or quality, more than likely their assumption, again, IF they knew what it meant in the first place would/should/could be that it is a 'better' (better, meaning more tests, which while this may not be the case with every product, is the case for the vast majority imo) product. How anyone who has an inkling of what build/version numbers mean could assume that taking that many tries to get it right is a bad thing is what's beyond me.

2006-06-28 Reply Admin

Anonymous:
Anonymous:
Well from a marketing perspective there is a point to what these guys are saying.

No, there isn't. I defy you to find me a single user who ever came to this conclusion. Come to think of it, I defy you to find more than a handful of users who even know what/where the version number is usually stored, and who even notice its existence.

Anonymous:
I don't consider this one a genuine WTF. This is just marketing and tech miscommunicating, nothing really stupid, both sides have a point here.

Not even close. Marketing are working from a flawed and baseless position of "what if" they found a user stupid enough to object to version numbers, but only in the case of their product (or they'd still be using Windows 1.0, IE/Mozilla 1.0, etc, etc, etc). There is no actual problem here, but they're still wasting developer time making them jump through trivial hoops so it looks like they're involved, on-the-ball and Taking Charge.

Having been a marketing and product manager, I have to agree there is NO marketing perspective here and they do NOT have a point except that they are in charge of things they are not clue-enabled to handle. And, worse, the managements and marketing folks seem to hate their customers or think they're retarded -- THAT'S always a good sign . . . not -- so if you'd mind posting the company name, I'd like to short the stock!

2006-06-28 Reply Admin

emurphy:

Why didn't you just offer to generate (say) 574 dummy issues and mark them complete?

Thank you!! I feel like the only sane person here. When a solution is simple, OFFER IT! I've never written an application that starts at "1" for exactly the same reason. Is this a database app? Just modify the sequence to start at some random, large number: 7839289

Abusive? It's 5 seconds of work!

2006-06-28 Reply Admin

Anonymous:
I think they are right. He can always change version and build number to md5(version and build number).

It looks cool:

Version d41d8cd98f00b204e9800998ecf8427e

That's it...I give up

nobody on the face of this planet knows what a hash function is

snoofle · 2006-06-28 Reply Admin

I don't recall (could be mistaken) any post ever having gotten 200+ posts - we are *this* close!

To quote some immortal words:

Let's Do Iiiiiiiiiiiiiiiiiiiiiiittttttttttttttt !!!!!!!!!!!!!!!!!!!

Eric Shinn · 2006-06-28 Reply Admin

Anonymous:
Anonymous:
I think they are right. He can always change version and build number to md5(version and build number).

It looks cool:

Version d41d8cd98f00b204e9800998ecf8427e

That's it...I give up

nobody on the face of this planet knows what a hash function is

If you know all possible version strings, it would be feasible for the support team to brute-force that backwards when they need the original info (though inefficient)...

2006-06-28 Reply Admin

Eric Shinn:
If you know all possible version strings, it would be feasible for the support team to brute-force that backwards when they need the original info (though inefficient)...

I'm going to assume that you're being funny and not trying to give me a migraine, but on this forum its like 50/50 odds:

1) If you knew all the version strings, you wouldn't need to brute force the hash. You would just compute hashes for all version strings and cache them. Then compare the user's hash to your known hash and work backwards.
2) You're still disregarding the fact that two strings can have the same hash. In that case, you can only narrow it down to two possible versions.

Bus Raker · 2006-06-28 Reply Admin

Anonymous:
NT, as we all know, stands for No Troubles. So putting that in there would have been redundant.

I heard that Windows NT stood for 'Nice Try'

snoofle · 2006-06-28 Reply Admin

Bus Raker:

Anonymous:
NT, as we all know, stands for No Troubles. So putting that in there would have been redundant.

I heard that Windows NT stood for 'Nice Try'

Really? I heard it was for "Not Today"

Bus Raker · 2006-06-28 Reply Admin

Anonymous:

This is not a WTF. Managers and marketers aren't always idiots, but I think programmers are too quick to assume that they're always right.

We are always right, that's why we do the important stuff. Management/HR sits around and determines my hours and salary. Customer service always me to be more efficient by consolidating the incessant whining of the customers into problems for me to fix. Accounting makes sure I get paid and figures out what to charge for my products.

I am the center of the universe: the developer. And it's all about me.

ME ME ME ME ME!

We at 200 posts yet?

2006-06-28 Reply Admin

"We never shipped a version x.0, "because x.0 releases are always buggy." So instead our first non-beta release was 1.1"

Who would buy a 1.1 release? They would know it was just a patch slapped on buggy product.

At the startup where I work we started our product versions at 2.5 and work up from there.

Our marketing person made us take our names out of the about box, but taking out the version numbers is a gigantic WTF. The googles, they do nothing!

2006-06-28 Reply Admin

Sinedia:
Bus Raker:
If s/he had suggested something like 'let's use the release date to indentify the product instead' it might make some sense. I mean, who really has two releases in the same day?

We do... quite regularly actually :(

I've seen it happen too. I've been involved in building/controlling software to be used in integration & test in a highly dynamic, schedule-driven environment. (It was for space, so there's a serious deadline looming -- launch day. That gets booked years in advance so the booster can be procured, pad time obtained, launch support, coordination with other facilities on the same range, etc. Moving the launch date even for a commsat is very expensive.) We had to get our stuff done in time for the customer to get it integrated into the spacecraft and do all of their testing as well. It takes a lot of coordination.

Anyway, in that sort of an environment, you also need to have a lot of control. We can't have builds that will be confused for other builds on the manufacturing floor. If we do a release candidate build at 8 AM and it fails its checksum or blows a raspberry when they try to load it onto the target hardware, then we're going to have to do another build as quickly as humanly possible and make sure that although it's still a candidate for the same release (so still version 1.8.2 or whatever), they can tell it's not the 8AM build. We did that with the version number plus the compile time.

On more than one occasion, we actually released the morning build, then were ordered to build an afternoon one, and engineering found a way of reliably supporting that without compromising traceability or quality. I was rather proud of that. ;-)

2006-06-28 Reply Admin

Stupid people abound. If you say you're going to quit or resign, then you are just as stupid because you're going to find the idiots cousin at your next job. Adapt. Overcome.

Ah, the refrain of bitter jaded people who have never been at a decent company. I'm sorry, dude.

rbriem · 2006-06-28 Reply Admin

emurphy:

Why didn't you just offer to generate (say) 574 dummy issues and mark them complete?

Better still, don't give the customers the ticket numbers! We can't have them knowing how many calls we get now, can we?

I worked on one project where the company guaranteed - in the contract - that the software version numbers would never go into double digits ... there would never be, for example, a 1.12, or a 12.12. They started numbering at 1.1 and within three months they were at version 4.6.9.3.4 ...

2006-06-28 Reply Admin

Anonymous:
So my XP version is Build 2600.xpsp_sp2_gdr.050301-1519

I checked with Microsoft, they did change XP 260,020,503,011,519 times.

Thanks - Just got ni trouble for laughing to loud!

2006-06-28 Reply Admin

We did 62 builds yesterday, our product must really suck.

Of course every 30 seconds we check to see if new code has been checked in, build that code, run unit tests on that code, run automated regression tests on that code and then make it immediately available for the QA should they choose to use it.

I guess I need to go back to 1 build every 2 or 3 months...

2006-06-28 Reply Admin

Sindri:
This would actually also make a good Dilbert Strip.

You're sure this isn't lifted from some comic. It's more than ridiculous enough

2006-06-28 Reply Admin

Anonymous:
"Please have your team remove this information in our next release."

Good thing that's probably the last request they are ever going to make concerning versions. After all, without release numbers, how could they properly identify the *next* release? :)

After all, theres no next release...

belef

One Version to Rule Them All

Leave a comment on “One Version to Rule Them All”