clickdeck360’s posterous

 
« Back to blog

The Mythical Man Weekend

The "I can do it in a weekend" beast reared its naive head on Hacker News recently which in turn generated some intelligent discussion on the topic. We've all been there and we've all learned the hard way. It's amusing to watch others step on that rake as you once did. The topic touched a nerve for me. We are building ClickDeck360 to help developers and PMs avoid the clusterfuck involved in building products. We hope that silly ideas like "I can do that in a weekend" get analyzed, discussed and tested out before a line of code is even written. We hope that people spend their weekends with family instead of working because someone said "Oh that, that's a 4 hour feature".


Let's run through a quick exercise:

Initial comment: "I can build StackOverflow in a weekend." 

Oh really? Draw it for me, on a napkin.



Whatcha got there? 
"Questions, each one with a vote count and some answers. Timestamp, username, number of answers."

Ok, maybe you can pull that off in a weekend...  Oh so do people have to log in to ask these questions and vote?
"Yes."
 
Ok, so let's add a login system in here. Don't forget the password reminder mechanism and the emails. What about the edge cases? You know when someone tries to vote but they are not logged in? What's the flow there? Ok let's spec that out too. Are you going to code it from scratch? No? You'll use Django's password system? Ok, so will you be spending some time integrating that?



Ok. Now it's good. Wait, how's search going to work? Do we index every word? How about tags? Ok, let's do tags. So the author will tag the question. Edge case alert: can tags have spaces? What happens if the author types in the same tag twice? Do we parse out duplicates from the tag list? What happens when the user clicks on a tag? Do we resort the results page? What order?

How about nesting? Can I respond to a commenter? How many levels of nesting? How will the UI handle that? Will it be too hard to navigate? Ok, then should we just leave comments on each response in a flat list?

Really, I'm just scratching the surface here and asking the questions that first come to mind. But I'll ask again: does this look like a "weekend's worth" of work, or a weekend's worth of questions?

Comments (6)

Jul 06, 2009
Brenton Simpson said...
Build the prototype in a weekend. Show it to people - get their reactions.

The goal of the weekend project is to make failure cheap. Get the core functionality working and play with it. If it works, keep going. If not, more on.

It's certainly helpful to break down a project into features, but it also helps to validate your concept before you start worrying about things like duplicate tags.

Jul 06, 2009
Dennis Gorelik said...
I agree with Brenton that weekend project is just a prototype [usually hardly working :-)].
Production-quality project implementation usually takes at least several months and usually years of multiple iterations.

One of hard-to-solve problems with mass-audience web projects (like Stack Overflow) is dealing with spam.
That definitely cannot be solved on a week-end -- I know it from my experience with www.postjobfree.com

Jul 06, 2009
ststrat said...
You haven't made failure cheap if it keeps costingyou your weekends, you've made it far too dear.
Jul 07, 2009
rosejn said...
I can't tell if this is just marketing hype for clickdeck or a genuine post. Weekend hacks are wonderful, and my guess is that a number of great products and projects have started as weekend experiments. No, of course you don't create the next Stack Overflow in a weekend, but if you wanted to try a new variation on the same idea, say using a different style of voting or something, then a quick and dirty weekend version might be just the thing. All of the use cases and documentation and stories and whatever other process side-effects might get thrown onto ClickDeck are really a waste of time if the core concept they are based around sucks. Double entry tag parsing? I don't think that should be on anyones list until the whole idea of the site has been explored and experimented with. Besides, not everything in life is about creating products. A lot of weekend hacks are purely for the joy of satisfying the programmers curiosity, and that would be a pitty to waste.
Jul 08, 2009
dogslovesalsa said...
"Mythical Man weekend" compared to my "Actual Man weekend" never stacks up. I did a quick comparison: http://www.semanticinternets.com/2009/07/mythical-man-weekend/
Jul 08, 2009
cannikin said...
You can create apps in a weekend: http://r09.railsrumble.com/ I think the problem comes in when you say you can duplicate some existing site, which has been battle-hardened in production for a while, and don't consider all the annoying bits like performance, database indexes, performance, scaling, architecture...all that stuff that gets in the way of actually building the app. :)

Leave a comment...

 
To leave a comment on this posterous, please login by clicking one of the following.
Posterous-login     twitter