The Daily WTF

Tales from the Interview: The Missing Interview, Infantile Expectancies, & More

Alex Papadimoulis — Tue, 09 Feb 2010 14:00:00 GMT

The Missing Interview (from Charles Ross)
I went for an interview to work as a junior IT support Engineer at a certain Royal bank here in Scotland. It was a late interview, around 4:45 in the afternoon, and I turned up at 4:30, sharply dressed, and with all the documents I'd been requested to bring. Since this was a bank and security was a must, I had a full five year history sitting in front of me.

I sat down and was quickly ushered into an interview room. I sat there for 20 minutes waiting, occasionally sticking my nose out the room to see if anyone was coming. After another ive or so minutes, it was about 4:55 and I decided to go hunting for someone.

Eventually, I found a receptionist getting ready to leave, and asked her where my interviewer was.

"Oh dear," she responded, "she forgot you. Right, go back to the room and fill these in."

She handed me a pile of papers and ushered me back to the room. Here, I got my first look at the papers...it was the forms the interviewer fills her opinions in.

At this point, I filled the form in saying how brilliant I was (literally saying "Candidate is brilliant! Hire Him!" for every field), gave it to the receptionist and walked out.

Sufficed to say, I didn't hear back.

Infantile Expectancies(from Roger Garrett)
I'm currently retired, but am always on the lookout for interesting ventures. I recently came across a job posting for a start-up company offering equity in exchange for work. It looked intriguing so I wrote to them.

John, the apparent CEO, responded.

Thanx for your reply. The aim is to produce a software vehicle that 
provides automated, simultaneous text translation for the major 
languages of the world.  For example, a business with an existing 
website signing up for this service would have their entire website 
translated into the texts of the major world languages, thereby 
increasing their accessibility multifold.  The service would be a 
monthly fee-based service.

I was a little confused about the model. In addition to the enormous technical challenges, automated translation is medicore at best and is already free with Google Translate and browser plug-ins. I wrote back to John asking for some clarification. He seemed a bit irked that I didn't sign on right away.

The fact is, from a programming standpoint, we really need a 
supergenius ability level of programmer for this project.
 
From a marketing perspective, we are not inviting input or 
commentary from any potential programmers. In fact, for any 
potential programmer to harbor qualms, doubts, and/or opinions 
regarding the marketing aspect of this project would unneccesarily 
belabor and cloud an already daunting task.  If we wanted marketing 
advice or personnel we would advertise for such.

Wait... what? I suppose I could have just let it go, but I couldn't help myself replying with "If you are so insistent on not addressing questions raised by a potential partner, and that's what you're asking for, a partner, then you're not going to find anyone who would abide by such absurdities." He responded shortly thereafter.

Opinions are like assholes, Roger, everyone has one.  Yours is now 
unwelcome.  So cease and fucking desist.
  
Contrary to your "opinion", the only absurdity I've encountered so 
far is your assumptive, ignorant and irritative attitude.  I do 
agree to a partial extent to one thing you said, we can be assured 
that we won't find "you" as a programmer.  In fact, I doubt you 
will find many companies willing to tolerate your infantile 
expectancies.
 
Good luck with that Roger, you will surely need it.

Not surprisingly, I have yet to see the "company" produce a product. Though, I do see an ad on craigslist for a start-up company offering equity in exchange for work...

Someone Like Kevin (from Jeanne)
A recruiter emailed this job description the other day:

NEEDS: 
 - a quick learner (ramp up on simulation software, GUI, API 
   interfaces, distributed applications) 
 - a clear communicator, expresses self well at all levels, 
   especially in writing (writing up bugs), writing test cases, 
   collaborating w/ development engineers 
 - well versed on s/w QA practices, QA methodology, QA process 
 - Scripting, shell, automation tools, test tools 
 - someone like Kevin 
 - 3-5 years experience 
 - prior software development experience 
 - Bachelors' in Comp. Sci or Math minimum, with the computer 
   programming or QA work experience.

I quickly decided it wasn't for me. I'm nothing like Kevin!

What do you mean?(from Casey)
Part of the interview process at my company involves writing a sample application in Java and sending it into the team for review. People who do well on the assignment are brought in for the face-to-face interviews, and each member of the development team will sit down with them one-on-one. One of the guys we brought in bore a striking resemblence, both in appearance and mannerism, to Chris Farley's "motivational speaker" character on Saturday Night Live.

The first red flag came up when he spent the first half hour of the interview bad-mouthing his current employer.

Then the second red flag came up: he'd spent over 20 hours on he assignment. His UI alone took 12 hours, and the bulk of that time was spent on a tabbed panel (which should have been trivial). And then I asked him why the print option wasn't implemented.

"Well, that was on the second page of the assignment," he said.

I wasn't quite sure of the revelvance of that, so I asked him how he would implement a print option.

He got a confused look, "What do you mean?"

I tried to get more specific and said, "Using Java, if you wanted to send data to a printer, how would you do that?"

"I just don't get what you're saying," he said, growing more irritated and confused by the moment.

I was running out of ways to phrase it. "It's not a trick question. Java. Printing. Anything come to mind?"

He looked at me blankly.

Finally, I had lost patience. "Ok, you have text on your screen. You want to get that text from your screen onto a piece of paper. You're writing a program in Java. How do you get the text from your screen onto a piece of paper?"

His brow furrowed and his gaze grew distant. I was beginning to think that I smelled something burning, when he weakly supplied "...the printing API?"

My relief was palpable. "Good! Yes, the printing API! That is how you print in Java."

The interview ended shortly after that. When I compared notes with the other interviewers, it was pretty unanimous: he's a talker that goes for volume of information rather than accuracy. Plus, on his way out, he complained that the interview had taken too long.

CodeSOD: Phishing for a Refund

Mark Bowytz — Mon, 08 Feb 2010 14:00:00 GMT

As I'm sure is the case with many of you, I sure do love me some tax refund. Once my W2's and other year-end tax forms documents come in the mail, I get my e-File in and wait for my refund to be direct deposited.

Now, since the whole refund process involves computers and the internet, of course, it's a prime target for spammers and phishers who want nothing more than to ignite a little FUD and get some of your hard earned cash.

The below snippet comes courtesy of Andy F. who writes, "Idiotic spammers and phishers probably provide your site with a glut of terrible code, but I got a chuckle out of this credit card validation function found in an HTML file attached to a 'UK Department of Energy Refund Notice.' It also included a variety of non-valid PIN number checks."


function Validate() {
   if (document.logonForm.my_card.value == "") { 
      alert("Please enter your card number !"); 
      document.logonForm.my_card.focus();
      return false; 
   }

   if (document.logonForm.my_card.value == 0000000000000000) 
      { alert("Invalid card number."); return false; }

   if (document.logonForm.my_card.value == 1111111111111111)
      { alert("Invalid card number."); return false; }

   if (document.logonForm.my_card.value == 1112223344443231)
      { alert("Invalid card number."); return false; }

   if (document.logonForm.my_card.value == 2222222222222222)
      { alert("Invalid card number."); return false; }

   if (document.logonForm.my_card.value == 3333333333333333)
      { alert("Invalid card number."); return false; }

   if (document.logonForm.my_card.value == 4444444444444444)
      { alert("Invalid card number."); return false; }
...
}

A pretty WTF bit of code to be sure, but Andy's last sentence got my gears turning - Could there really be...MORE?! So, after scanning Google, I found that it was a piece of hot debate and included the original web page in its full glory.

As it turns out dear readers, the above code snippet is merely the tip of the iceberg. The entire file itself is the quintessence of WTF. Of course, I never expected that a phishing site would be a beau ideal of good coding practices, but...ugh...you really have to view the source to see what I mean.

Here's a link to the source - as a text file

However, for the adventurous - here's a link to a local copy of the page so, preserved as it was originally.

Sponsor Appreciation, Banzai Bouncer, Untraditional Data Rack, & More

Alex Papadimoulis — Fri, 05 Feb 2010 14:00:00 GMT

Please show your support for The Daily WTF by checking out the companies that have been kind enough to sponsor us. And, in doing so, I’m sure you’ll find some pretty cool products and services built by like-minded developers and IT professionals.

The Daily WTF Sponsors

Microsoft WebsiteSpark - a great program for web shops and freelance web developers and designers where you get some great software (Visual Studio Pro, SQL Server, Server 2008, etc), at no upfront cost for three years; it also provides support and resources to help grow business

Peer 1 - provides award-winning Managed Hosting, Dedicated Hosting, Co-location, and Network services offered through 15 data center across North America. With over 10,000 businesses hosted on their legendary SuperNetwork™backbone, PEER 1 delivers one of the highest server performance and network outputs in the industry.

MindFusion - a great source for flow-charting and diagramming components for a variety of platforms including .NET, WPF, ActiveX and Swing

SoftLayer - serious hosting provider with datacenters in three cities (Dallas, Seattle, DC) that has plans designed to scale from a single, dedicated server to your own virtual data center (complete with racks and all)

SlickEdit - makers of that very-impressive code editor and some pretty neat Eclipse and VisualStudio.NET tools and add-ins, some of which (Gadgets) are free. Check out this short video highlighting just one of SlickEdit's Visual Studio integration features.

And now, back to our regularly, completely off-topic scheduled program.

"I came across the Banzai Jump 'n Go Obstacle Course Bouncer on Walmart.com," notes D. Kablanc, "it looks like Walmart needs to hire a new Photoshop artist."

"Don't they normally put those disclaimers in the fine print?" wonders Martin Sapsed, "just seems like a strange place putting it next to the sign promoting the water park."

John Spalding writes, "that sure is one Heavy Duty traffic cone."

"Wait a sec," Rowan Pope writes, "I thought cars came with all DVD players?"

"Ugh," an anonymous reader wrote, "the local newspapers really are't catching up..."

"I was in an Ethnic grocery shop in Glasgow," David writes, "and spotted this rather untraditional use for a 42U data cabinet."

"I saw this in a Japanese restaurant in Lisbon," Miguel Lourenço noted, "do fish really say grrr?"

"Ah, the EULA Hotel," writes Samuel Hartman, "I'd hate to see their reservation form!"

"I remember this one," Justinas Lelys writes, "it's from when Mickey Rourke is female and Winona Ryder male."

"I can't say I was surprised to see that I'm now being charged a fee for being charged fees," writes Lincoln, "and with the wierd alignment, I have to wonder if they're saving room for the FEE FEE FEE."

Emergency Faxes

Mark Bowytz — Thu, 04 Feb 2010 14:00:00 GMT

As far as technologies go, faxing is ancient. It predates the telephone by over a decade and, despite vast advances in scanning and email technology, the fax still remains a standard form of communication.

When a transmission goes out, the occasional telecommunication ‘hiccup’ or line noise can corrupt the fax. Most modern fax machines have some rudimentary error handling that will alerts the user that the fax should be resent.

This method of managing failures has been working so well that no one really saw fit to change it. That is, until a business analyst at Torre L’s company had a bright idea.

“What if the user isn’t babysitting the fax?" he worried. "What if the sender’s fax machine can’t detect that there was a problem? We should fax back an error report!”

Though the analyst's concern was valid, Torre and his group argued that sending back a failed fax transmission would not necessarily make it easier.

They argued that it could tie up the original sender's machine and could actually prevent someone from resending their original fax.

"It's okay," the analyst replied, "Our software can send and receive in parallel. This idea is the best thing since the iPod! It's bulletproof!"

But the arguments were for nothing; in management's eyes, the business analysts were always right and the feature was completed in short order. Named the "FaxBack" program, it did exactly as the name implied; it faxed a failed transmission back to the sender (as identified by Caller ID) moments after the failure occurred.

Everything went along without a hitch for a good long while. That is, until two police cruisers — sirens blazing — showed up at Torre’s office. They said they were responding to a 911 emergency call, but there was no emergency and nobody claimed to have dialed 911.

They officers left shortly thereafter, but, in the wee hours of the next morning, two different police cruisers rushed to the scene. Again, there was no emergency and no 911 call, so they too left without incident.

But the third time, when a cruiser showed up the following afternoon, the officers refused to leave until the source of the “disturbance” was identified. The police department was certain that a call came from a phone number within the company, and they demanded to know what exactly was going on.

After spending the rest of the afternoon and part of the evening of tracing phone calls internally, Torre traced the 911 calls to the data center, or more specifically, to FaxBack.

The faxes machines, like the office phones, had to dial "9" for an outside line. So when FaxBack responded to a failed fax from New Delhi by dialing "9", and then India's international country code of 11, a "special case" rule in the the telecom system kicked in.

Whoever had first installed the telecom system had the bright idea to treat “911” as a special case since, ordinarily, one might not think to first dial “9” in case of an emergency. The special case applied to the lines in the fax pool -- even if the caller would only ever be a computer.

CodeSOD: Almost Counting Past 100

Mark Bowytz — Wed, 03 Feb 2010 14:00:00 GMT

"At the contract shop where I work," writes John S., "I have been assigned to a new web-enabled mapping program to help take a look at some of the issues they've been having."

"When an item is added to a map, it is given a label, such as Item #1, Item #2, etc., with the number on the label incrementing for each new item. We had been having a problem where map labels were not being assigned uniquely when there were more than 100 items per map. It was always starting at Item #100 when reloading the map from the database. This was causing issues since it was the map label name that was being used for the unique identifier (don't get me started on that). Curious, I took a look at the code to decipher how the label ids were being assigned."

setCount: function(label) { 

        var layer; 

        var count; 

        if (label.charAt(label.length - 2) == ' ' ) { 

                layer = label.substring(0, label.length - 2); 

                count = label.substring(label.length -1) * 1; 

        } else { 

                layer = label.substring(0, label.length - 3); 

                count = label.substring(label.length - 2) * 1; 

        } 

        if (this[layer] < count) { 

                this[layer] = count ; 

        } 

}

"Each time the map was loaded, a function iterated each layer to find the items assigned to that layer, and then to parse out the number from the code and keep track of the maximum id. Apparently they had the foresight to account for more than 10 items, but not for more than 100."

Passed Around

Lorne Kates — Tue, 02 Feb 2010 14:00:00 GMT

The rejection had taken three months to arrive, and now somebody, somewhere, owed Luis K an explanation.

Why had a required feature been rejected? He couldn't tell from the cryptic jumble of control codes and received/forwarded stamps that overflowed the "office use only" box. The internal trouble-ticket system just showed "handled externally".

He'd discovered the missing feature his first day with Hausdorff Solutions. He'd received a satisfying "good catch" after filing the feature request with his department head. Surely, she'd know why.

Passed On

"I have no clue," Shirley stated, sliding the paper back to him. "We don't handle these sorts of features."

"But we do navigation functions," Luis said.

"This," she replied, pointing a stiff finger at the form, "Is a mapping function. Mappings are a completely different department. I have no idea how they handle these things. You should submit a 27b-6 for clarification."

Luis blinked, stymied at the disconnect. He looked at Shirley's accusatory finger.

Sometimes all the left hand needs to know is where the right hand is, so it knows where to point the blame.

"I'd hate if anyone got in trouble for missing this," he said, frowning. "Who did this get passed to, so I can clarify it with them?"

"Vern in Accounting," she offered quickly, pointing out his "received" stamp.

Vern couldn’t have made a price quote without an estimate from Lena in UI Design, and she couldn't modify anything without input from Paul in User Acceptance. Anything Paul did first got submitted for approval by Mr. Fischer, the project lead.

Mr. Fischer, his face awash in the azure glow of his monitor, didn't even look at the offending slip of paper.

"We do not deal with new features,” gruffed the man in the always-pressed brown suit.

"It isn't new, sir!" Luis pleaded. "It's in the specs!"

"Would it need to be added to the existing navigation subsystem?" Mr. Fischer asked.

"Yes, but—"

"Ergo, it is a new feature. Not our company's responsibility. You'll need to file a 27b-6 with The Consortium."

Luis paled at the thought of tangling with that bureaucratic kudzu.

Passed Up

Hausdorff and three other companies formed one of the project's action groups. Their group, along with three other groups, were the German conglomerate. The conglomerate was sub-contracted to another conglomerate of European companies, who were themselves sub-contractors of an organization of customers. Those customers, along with their North American and Asian branches, formed The Consortium. And each member company in The Consortium could, in turn, sub-contract their work out to other companies— fractally on and on.

"No, I don't have a 27b-6," he said to the young woman on the other end of his fifth extra-office phone call. "I just want to know who the request was passed to, so I can find out who they passed it to— and so forth."

"I can find out for you, I suppose— if you really feel you need to do that."

Luis taken aback by the tone of her voice— resigned rather than disinterested. "Unless there's a better way?"

"Perhaps," she replied, her voice hush. Luis could picture her hunched over the phone, a hand over the receiver as not to be overheard. "But if you tell anyone, I'll deny that we spoke. You didn't phone through official channels so you can't prove anything!"

"I won't tell," he said, glancing nervously around his cubicle.

"The truth is, there are only two companies qualified to handle any request; the one that generated it, and the one that handles it in the end. Everyone else is just an intermediary. Well, maybe those two companies, since they work in the same field, know each other already, right?"

"Then maybe," Luis pondered aloud, following her line of thought, "the process could be streamlined a bit by establishing a direct link between them, don't you think?"

"I would never!" she retorted, aloud and deliberately. "Good day, sir!"

Luis hung up, and stared at the form. Nothing looked familiar, just an overlapping litter of foreign control codes and forwarding stamps.

And then, the form stared back into him.

Passed Back

Luis recognized one of Hausdorff's sub-contractor's control codes. Their peculiar format of encoding the date as a hex number was unmistakable. Had he filed the request directly with them, half of the pass-throughs could have been eliminated.

And then he recognized the mark of one of that company's sub-contractors...

Having seen the unseen, he mentally sliced section after section from the jumbled chain. That left just three nodes; Hausdorff to Lindenmayer, Inc. for feasibility, Lindenmayer to Banach Systems for cost analysis, and then...

Back to Hausdorff for approval?

Furious, Luis opened the internal ticket-tracking system, and entered Banach's tracking code. If this could have been solved with an intra-company request, or worse, a face-to-face...

An entry from one and a half months ago popped up.

"This is a new feature request. We only deal with existing features. Submitter must file a 27b-6. Status: Rejected by Luis K."

After a few moments, Luis swept the fragments of his mind off his desk, and dutifully filled in a 27b-6.

CodeSOD: Else... where?

Alex Papadimoulis — Mon, 01 Feb 2010 14:00:00 GMT

"I had a professor once who said that given enough NAND gates, he could rule the world," writes Rob B. "This was a roundabout way of saying that, using a whole bunch of NAND gates, you could create the function of any other logic gate. You shouldn't, because the other logic gates exist and it would be hugely wasteful to use NAND gates to do the same thing, but it can be done. It turns out this applies to code as well."

"We got some utterly garbage C++ code from a subcontractor. The error-to-lines ratio was amazingly high, and there were a lot of things to hate about it (like having one global function to get bits from a binary value which didn't work, and several different localized one-off solutions which did work). My main WTF moment, however, was the following."

while(true)
{
    if(mainType == 7)
    {
        subType = 4;
        break;
    }

    if(mainType == 9)
    {
        subType = 6;
        break;
    }

    if(mainType == 11)
    {
        subType = 9;
        break;
    }

    break;
}

"Just look at that for a minute," Rob continues, "I spent a lot of time thinking about this. The while loop is only there so that the break can be used to jump over code. I've never seen an unconditional break at the end of a loop before, but there's a first time for everything."

Rob added, "all I can figure is that the developer honestly didn't know that there is such a thing as 'else if'. But he did know about 'if', 'while', and 'break', so he cobbled together an 'else if' in the most ridiculous way possible."

Error'd: Oh, I see.

Alex Papadimoulis — Fri, 29 Jan 2010 14:00:00 GMT

"I'm pretty sure that the price tagging software at Forever 21 trims the beginning and end of the Description field," George. "Either that, or these are more popular than I realized."

"The new grading software the school uses takes security very seriously," Ben W-R writes, "not even the intended recipient can decipher the progress reports."

"I think only Kurt Godel can solve this one," writes Paul G.

Frank got this exceptional note on his invoice.

"I mistyped a date on some web form," Michael Sondergaard noted, "obviously, it wasn't wrong enough."

"I got this error when trying to launch Battlefield 2," notes Russ.

Phillip Sitbon commented, "I love knowing when software points out bad self-referential habits!"

"I've been planning the migration of a database server to a new Veritas Cluster Server-managed cluster and thought that I should verify the install procedure," Jarrod Spiga writes, "but let's just say the instructions supplied by Veritas/Symantec are less than clear, especially on pages 123 and 124."

"At the visitors' center in the Little Farm in Tilden Park (Oakland, CA)," notes Colin, "there's a scale that tells you your body's water volume. Mine is apparently '-STACK @ 555'."

"I've been using a Yamaha AW4416 multitrack recorder a lot lately," noted Erno Palonheimo, "this is what happens if you don't have a hard disk hooked up and try to enter the song edit mode. Extra points for the... creature. Penguin? Maybe."

Classic WTF: The Network Batch File Virus

Alex Papadimoulis — Thu, 28 Jan 2010 14:00:00 GMT

The Network Batch File Virus was originally published on March 15, 2007.

The early 90s were exciting. Tim Berners-Lee invented HTML and created the first of the many internets we have today. A bunch of dancing dudes in foil costumes built the first Pentium processor. And who can forget Eritrea gaining independence from Ethiopia? Well, I could, but I wasn't following Ethiopian politics so much those days.

At the time, Chilton W. was working at a company that provided software that auto insurance companies used to keep track of users, policies, cars, etc. Things were going great, too. They had sold their software to one third of all auto insurance providers in the U.S.

The architecture consisted of a main system (which they owned) at each insurance provider's location, physically linked to other systems on a local network. This didn't work under the typical client/server model, though, it was a daisy-chained, peer-to-peer setup.

Often just one computer on the network could connect to the internet via dialup, so troubleshooting was a matter of dialing into the main system, checking for problems, then walking a user through the steps of testing each networked computer. It was usually the low man on the totem pole at the client's site that would get tied up for a few hours, changing configuration settings and testing. It was a tremendously tedious process, and the answer was generally either a bad cable, bad software install, or a dead hard drive.

Chilton saw an opportunity to simplify the process by deploying a remote diagnostics application. Deployment was, in fact, pretty easy — a batch file could upload the utility to the insurance provider's main system, then someone there could copy the utility to a floppy and manually get it out across all the other systems. Chilton still didn't like that he'd have to explain to each user how to copy the utility, though, so he added the ability for the utility to replicate itself across the network.

Chilton's script was a godsend. He could easily diagnose problems with one call, one batch file upload, and one execution of the utility. A few minutes later, a log file would be created that could easily and accurately identify the problem. Chilton's productivity skyrocketed to the point that he could solve ten or more problems in the same time it took his coworkers to diagnose one.

The future was bright for Chilton. He began showing others how to use the utility, and everyone loved it. When a relative offered to send him back to finish college, though, he put in his two weeks' notice. His boss requested that he write documentation about the utility, so he printed up the batch script and wrote extensive documentation about how it worked. A few days later, he left.

A year later, he ran into one of his old coworkers and asked how things had been going at the company. "A month after you left, everything went crazy," he said. "Management said we had to wipe all of our hard drives and recall our systems from the field. Tech support was down for months!"

"What?" asked Chilton. "Why?"

One of the comments Chilton had left in his script read "this self-replicates, similar to how a virus works." It was read by the wrong person, and a "security expert" was immediately brought in. His conclusion was that this was clearly viral batch file, undetectable by every antivirus software product on the market. The fear that an elusive virus had permeated throughout the network lead to the company hiring very expensive investigators to look into the issue. The head investigator found nothing, though, so he was fired and the investigation ended. Ultimately, new computers were bought for everyone, and tech support was able to get back on track.

CodeSOD: Break Out

Alex Papadimoulis — Wed, 27 Jan 2010 14:00:00 GMT

“I’m continually amazed by the unique and clever solutions developed by my colleagues,” Mark writes. “And I should say, I don’t mean ‘amazed’ in a good, innovative-idea-to-save-company-money sort of way. It’s more a wow, that’s more wrong than I could have ever imagined sort of way.”

“Take, for example, this snippet of code that I found recently.”

    for (int i = 0; i < dtModules.Rows.Count; i++)
    {
        if (dtModules.Rows[i]["id"].ToString() == ID.ToString())
        {
            dtModules.Rows.RemoveAt(i);
            i = 34598;  //used to jump out of the for
        }
    }

Mark adds, “perhaps the original coder thought that the word ‘break’ must be offensive to someone out? And really, it’s easier to just pick a random number to force the for-loop to exit and then hope that nobody ever has more than 34,598 modules.”

Nobody Does Business on December 31st!

Mark Bowytz — Tue, 26 Jan 2010 14:00:00 GMT

Steve's phone gave its distinctive internal ring.

"Steve! Hey! Happy New Year, man! Jeff here from Corporate AR!" the caller was speaking a mile a minute. "I sent you a critical email. Did you get it yet?"

"Nothing yet, but let me refresh Outlook," Steve S. clicked Send-and-Receive and waited for a moment. "Okay... yeah, I see it. Go ahead."

Steve gave the caller points for trying to hide his panic, but as things went on, it was quite obvious that something had hit some fan... somewhere. But for the most part, Steve mostly tuned him out. It wasn't the first "Jeff from AR" that he had spoken to.

Now make no mistake about it: Steve isn't an accountant, nor does he work with any financial systems on a regular basis. But the system that he maintains does, however, feed the corporate Account Receivable (AR) and General Ledger (GL) systems on a regular basis. Periodically, the bean counters reconcile the dollars in the AR system to what has been sent to the GL system as a sort of check-and-balance process and, almost always, everything is in sync. Almost.

In the first few days of January, it's the same song and dance. An email will land in Steve's inbox, and the source of the email is always who started within the past year or so. The subject is typically something ominous that makes the hair on the back of your neck stand up — **URGENT! CRITICAL ISSUE!** Billing System Clearing Account #12345678 for Business Units XYZ & ABC — and the email always comes complete with a few screenshots from systems that look as strange and foreign to a non-finance person as SQL would look to the person sending the request. However, the panic-worthy bottom line is always the same.

Jeff from AR continued, "the outage for Business Unit ABC is off by $1.3 million and I have no idea what is causing this. Can you please look into these items ASAP and let me know your thoughts?"

Of course, any time that a balance is over $1 million off, it tends to make anybody sit up and take notice. But for Steve, the feeling of déjà vu was quite palpable.

Did You Check the Logs?

Steve started into his research by asking, “When was the last time that GL was balanced against AR? And was it in balance?” While he never got a direct answer to that question, Steve eventually found out that everything seemed to be good right up until the end of December. “Were there any errors in the processing of the last GL file we sent you?” Steve asked, “ You know, the file that has the records with the effective dates of December 31st?”

“Not sure – I’ll have to check the processing logs and call you back” replied Jeff.

Moments later, Jeff called back and reported that the processing logs said that the problem was, Batch Consists of Posting Transactions in Different GL Periods. He snarkily added, "well, of course! That explains it. How could we be so stupid!?"

Steve thought the problem sounded familiar, so he called up the "GL Gurus" to look more closely at the problem and figure out what it meant. This is when some lights of recognition start to glow more brightly in peoples’ heads. “Hmmmm, it seems we’ve seen this before. What does it mean again?”

After lengthy debate, one of the gurus proclaimed “Oh, yeah, we can just change the dates in that file from December 31st to December 30th and then rerun the file.” So later on, lo and behold, the file processed cleanly and matched the GL 100%. Everyone was happy, vowed to remember to do that next year, and life moved on.

The 13th Period

This year, Steve simply had enough. He knew he wouldn't remember the issue the following year, and he was not content to just let the matter die.

Rather than sweep the matter under the rug again, Steve started doing some snooping on his own. Eventually, he found someone who had archived an email from years past.

---------- Forwarded message ----------
From: Lewis Dodgson <ldodgson@initech.com>
To: *GL-All <gl-all@initech.com>
Date: Wed, 2 Jan 2002 13:44:08 -0500
Subject: reminder—do not use December 31 as journal entry date

All,
It has been noted that we have received a couple of journal 
entries with a December 31 effective date. This is a reminder 
that using that date will post the journal entry to Period 13. 
Period 13 is reserved for business unit collapsing entries 
and Corporate year end entries.

Regards,
Lewis Dodgson
Assistant Vice President, Accounting & Operations

Amazingly, the GL system has special logic built into it based on the magic date of December 31st and it had been there for literally years. Steve wondered Didn’t anybody think that anybody processed transactions on the last day of the year? Why wasn’t this found out earlier? Sure there was turnover in the financial departments, but the gurus should have remembered the solution, after all, they were the GURUS for crying out loud!

Enough was now finally enough – Steve was now committed to do something, to keep this new year cluster&$^% from happening again. He weighed his options - Should he lobby for the dropping of the 13th period? How many systems would this affect? Could he make a case for such a project spanning different groups? Could he sell a techy reason for completely changing how the company did business to a bunch of stuffed shirts in finance?

In the end, rather than upset the corporate apple cart, Steve put in a simple change request so that his system won’t send 12/31/YYYY in any more GL files. If the system hits that date, it’ll use 12/30/YYYY instead. The GL system apparently doesn’t care. As long as as a date is “some time in December” and it’s not “December 31st”, things will be fine.

CodeSOD: Meaningless Progression

Mark Bowytz — Mon, 25 Jan 2010 14:00:00 GMT

"One of our desktop applications has a progress bar in it," Bryce N. writes, "and as I was working more and more with the code, I noticed that the progress bar would progress to a seemingly random part in the bar, but never past the halfway mark. This would probably be ignored, if it weren't for the fact that I noticed that my breakpoints would only be hit when the bar reached the 'random' mark."

"While I was trying to discover why, I found this in the code:"

     /* There is really NO eloquent way of calculating what the progress 
      * of a given method/task will be.  One task may be downloading or copying a file while another 
      * one might be grabbing huge chunks of data for file creation.  Since we want to see 
      * a progress indicator but can't determine this value, we'll simply play with it so it has the 
      * appearance of running; (i.e. we'll take it to 50%, execute the task, then come back and finish 
      * the progress upon completion.)  This is a Microsoft STANDARD... I'm sure of it! 
      */

     Random R = new Random(); 
     int _percentage = R.Next(Convert.ToInt32(.5 * _bar.Maximum));

     string status = ""; 
     for (int i = 0; i < _bar.Maximum; i++) 
     { 
         if (i == _percentage) 
         { 
             //execute the real code 
         }

         // Always perform progress step 
         _bar.PerformStep(); 
     }

"Well, there you have it, the code tells the progress bar to progress somewhere between 0 and 1/5 the length of the bar, then execute the real code. Brillant!"

Error'd: Scientific Phonation

Alex Papadimoulis — Fri, 22 Jan 2010 14:00:00 GMT

"Drat!" Arttu notes, "the limits of displaying long numbers caught me again!"

"It took a few tries," John writes, "but the second option worked."

Andrew writes, "I really hope I live to be more than 0 years old."

"IT professionals can wear many hats," Rich Lovely wrote, "glad to see our local college is preparing us for this one."

"Thanks for the self-esteem boost, CNN," writes Chris Heng.

"I finally matched the lady's head with the buff guy's body!" writes Chris H, "Does that mean I win?

"After a hard day's work debugging web services, I decided to visit the vending machine to treat myself" wrote Iain Collins. "Apparently fate has a sense of humor."

"I saw this today while browsing our University 'Cadetship Opportunities' notice board," Dean Camera writes, "Apparently not even the University staff read the darn thing."

"I stood there for a couple of seconds trying to figure out what this means," notes Jason Berberich, "I'm still not sure."

"I signed in remotely to help clean up the work computer over night," Jacob writes, "but then I came across this interesting problem with the disk-cleanup wizard."

Bad Code Offsets: Open Web Innovation

Alex Papadimoulis — Thu, 21 Jan 2010 14:00:00 GMT

By now I'm sure you've heard of the Bad Code Offsets project. I announced it here back in November and gave a pretty exciting update in December, where we were able to give a whole bunch of money to some great open source projects. But what was especially exciting was the The $500 Good Code Grant.

Tell us how your free and open source project prevents bad code from being created and show us how $500 would make a real difference in your project
— or —
Propose a new, free and open source project and show us how $500 would help you get it started

There were some fantastic projects that contacted us about the grant, but the response from Eric S. Raymond — as in, the Eric S. Raymond — made the best case for his project, GPSd:

We really do drive out bad code, in both direct and indirect ways, and we supply examples of good practice for emulation.

GPSD is a service daemon and device multiplexer that is the open-source world’s basic piece of infrastructure for communicating with GPS receivers, and it’s everywhere Linux is – running on PCs, on embedded systems, and on both OpenMoko and the entire line of Maemo cellphones. We’re directly relied on by dozens of applications, including pyGPS, Kismet, GPSdrive, gpeGPS, position, roadmap, roadnav, navit, viking, and gaia. If you’re doing anything with GPSes on an open-source operating system, GPSD is your indispensible tool.

...

The reporting protocols used by GPS sensors are a hideous mess — the kind of mess that tends to nucleate layers of bad code around it as programmers with insufficient domain knowledge try to compensate for the deficiencies at application level and wind up snarling themselves up in ever-nastier hairballs. Part of what GPSD does is firewall all this stuff away; we know everything about the mess so you don’t have to, and we present clean data on a well-known port in a well-documented wire format. We then provide client-side service libraries that will unpack GPS reports into native C, C++, Python, or Perl structures so you don’t even have to know about our wire format.

If our client applications had to deal with the back-end mess of poorly-specified NMEA 0183 and seventeen different vendor-specific binary protocols, I for dead certain guarantee that the total community bug load from GPS-related problems would go up by an order of magnitude. And I’d bet more than any of the $500 prizes the Alliance is offering on the bug count going up by two orders of magnitude.

We also try to drive out bad code indirectly in the same way we keep our defect level low — by providing an example of good practice that extends all the way up from our development habits to the zero-configuration design of the gpsd daemon.

We're proud to support the GPSd project, and will be disbursing the grant money right away. Ideally, we'd like to deliver a giant, over-sized check in person, but we felt the exorbitant production cost was not a good way to spend the Alliance's limited resources. And speaking of those resources...

Join the Alliance for Code Excellence!

Thank you to everyone who joined up as a Patron Member. As I mentioned before, The Alliance for Code Excellence is a not-for-profit venture that's 100% volunteer driven, and members pay for the hard costs (materials, legal, postage, etc) through their annual dues of $50.

So if you like what we do, feel free to become a member. A one-year membership is $50 and, as a welcoming gift, you'll receive a black T-shirt with our logo on it.

Become a Patron Member

Cost: $50.00
Select T-Shirt Size:

Good Code Grant 2.0

Thanks to our members' support, we'll be giving away another $500 to a small, lesser-known open source project that could use it effectively. The guidelines are almost the same as before: simply tell us how your free and open source project prevents bad code from being created and show us how $500 would make a real difference in your project, and we'll consider donating to you.

So what's different from last time? We're only considering existing projects, not to-be-created projects. It's not that we don't want to support ideas, it's just that we have another idea for those.

Announcing the Open Web Innovation Grant

We've noticed that there are a whole bunch of open source frameworks and applications, but there aren't too many free and open source websites that aren't inundated with ads. One great website that comes to mind is Ajaxload; it's a simple little website that makes those ajax "loading" icons for you.

We're not entirely sure why there's a lack of these websites around, but I suspect a large part has to do with the cost. Software tools aside, a website that generates a reasonable amount of traffic can cost a bit of money to run, and not too many people are eager to commit to spending that in the long term.

We're hoping to change that. The goal of the Open Web Innovation Grant is to remove the cost barriers to creating free and open source web applications. Here's what we'll do as part of the grant:

Web Hosting - from ASP.NET to Zend PHP, we'll get a prime space on the web for your project
Software Tools - if there's a particular IDE or control suite, we can get that, too
Advertising - we'll promote the project to developers through targeted advertising on popular developer sites like this one
Cash - up to $500 to get started

All you need to do is tell us your idea, how you would implement it, and what support you'd need to launch. Remember, the goal is to build a project that helps reduce the generation of bad code, even if that means saving developers time.

The 1,000 SLOC Offset

I was originally planning on ending things in the previous paragraph, but I couldn't resist sharing this. Bad Code Offsets can be purchased in all sorts of denominations, ranging from 1 SLOC to 10K. While we certainly expected the 1 SLOC's through 100 SLOC's to be sold, we were surprised when an anonymous supporter purchased a 1,000 SLOC Bad Code Offset. I'll leave you with a (crappy cellphone camera) picture of what the 1,000 SLOC offset looks like.

Unlike the lower denominations, these are oversized and come matted and framed in the color of your choice.

Bring Your Own Code: Avoiding the Splice

Alex Papadimoulis — Wed, 20 Jan 2010 14:00:00 GMT

One mistake that rookie carpenters will often make is to measure for trim molding – baseboard, casing, crown, etc – by the linear foot. Take the casing on a 7’ door, for example. Each leg of the door requires 84” of trim and the header needs 32”. If your house has 16 doors, and each side of the door needs 200” of trim, then that adds up to 533’ 4” (16 x 2 x 200”). And since you can get casing in 16’ boards, you’d need to order 34 boards to get the job done, right?

Not quite. You’d actually need 38 of those 16’ boards. Although each 16’ board can easily fit two 7’ door legs, the remaining 24” should be scrapped, as a splice in a header casing is about as professional as modHmm. And while door casings are relatively easy to measure for, baseboard and crown molding can get trickier.

A general rule of thumb – especially for stained wood – is that splices should be avoided at all (reasonable) costs. While it’s unreasonable to expect an extra-long board to be made for rooms longer than 16’, there’s no reason to splice a 7’ and 4’ section together. Actually, there is a reason, but that has less to do with materials cost and more to do with a lazy carpenter who estimated incorrectly.

Bring Your Own Code

Your exercise for the day: write a function that calculates the number of boards needed to trim out a house.

The input should be a series/array of integers that represent each individual measurement; for example, a single 7’ door would be { 84, 32, 84 }.
The output should be the minimum number of boards needed to complete the job
The board lengths avaiable can be one of three, with the longer boards prefered when possible:
- Easy - only 16’ boards
- Medium - 8’ or 16’
- Hard - an (additional) array of integers that is an input to the function

For bonus points, have your program print the cuts needed on each board.

Special thanks to the lazy carpenter that I hired (and all his spliced moldings) for inspiring this BYOC, and to Todd Lewis for encouraging me to finally write it up. If you have your own fun coding exercise, don’t hesitate to contact me!

The Daily WTF Sponsors
		Microsoft WebsiteSpark - a great program for web shops and freelance web developers and designers where you get some great software (Visual Studio Pro, SQL Server, Server 2008, etc), at no upfront cost for three years; it also provides support and resources to help grow business
		Peer 1 - provides award-winning Managed Hosting, Dedicated Hosting, Co-location, and Network services offered through 15 data center across North America. With over 10,000 businesses hosted on their legendary SuperNetwork™backbone, PEER 1 delivers one of the highest server performance and network outputs in the industry.
		MindFusion - a great source for flow-charting and diagramming components for a variety of platforms including .NET, WPF, ActiveX and Swing
		SoftLayer - serious hosting provider with datacenters in three cities (Dallas, Seattle, DC) that has plans designed to scale from a single, dedicated server to your own virtual data center (complete with racks and all)
		SlickEdit - makers of that very-impressive code editor and some pretty neat Eclipse and VisualStudio.NET tools and add-ins, some of which (Gadgets) are free. Check out this short video highlighting just one of SlickEdit's Visual Studio integration features.