Douglas Tarr

Azure Websockets 302 Error

Wed, 20 Mar 2013 16:04:12 -0700

If you see an azure websockets error where it falls back to long polling , you need to make sure you set your app to use .NET 4.5 in your web.config

Permalink | Leave a comment »

Using powershell 3 Invoke-RestMethod with Basic Authentication and JSON

Tue, 05 Mar 2013 12:05:24 -0800

This took me a remarkably long time to figure out. Enjoy!

$user = "myuserid"

$pass= "mypass"

$uri = "https://myresturi.com/api/something"

$json = "{

""foo"" : ""bar""

$secpasswd = ConvertTo-SecureString $user -AsPlainText -Force

$cred = New-Object System.Management.Automation.PSCredential ($pass, $secpasswd)

Invoke-RestMethod -Uri https://api.intercom.io/v1/users -Method Post -Credential $cred -ContentType "application/json" -Body $json

Permalink | Leave a comment »

Setting IPRestrictions in a Windows Azure Cloud App Deployment

Sun, 24 Feb 2013 17:56:00 -0800

To set IP Restrictions in Windows Azure, you have to do 2 things, create an azure startup task in your service configuration (.cscg file), and a cmd file that runs it.

Create a Service Configuration Startup Task to run a startup command

Create a startup command

This startup command checks the "RESTRICTIP" variable so that you can have different deployments that you can set whether or not to restrict IPs by (beta, test, production, etc). It also bypasses this if you are running the Azure Emulator.

This installs the IP Restriction Windows Role, unlocks the IIS Config, and sets the appropriate IP Restrictions (put your IP in place of 1.2.3.4)

if "%RESTRICTIP%"=="false" goto :EOF

if "%EMULATED%"=="true" goto :EOF

powershell Install-WindowsFeature Web-IP-Security

%windir%\system32\inetsrv\AppCmd.exe unlock config -section:system.webServer/security/ipSecurity

%windir%\system32\inetsrv\AppCmd.exe set config -section:system.webServer/security/ipSecurity /~ /commit:apphost

%windir%\system32\inetsrv\AppCmd.exe set config -section:system.webServer/security/ipSecurity /allowUnlisted:false /commit:apphost

%windir%\system32\inetsrv\AppCmd.exe set config -section:system.webServer/security/ipSecurity /+"[ipAddress='1.2.3.4',allowed='True']" /commit:apphost

This particular solution will not work well if you are trying to have IP restrictions on your staging deployment, and not your production deployment, since the startup command only runs on a deploy, and you will likely just swap VIPs to go from staging to production.

Permalink | Leave a comment »

Don't prototype on your own platform

Sat, 16 Feb 2013 13:48:56 -0800

When building a new feature for your website, avoid using the legacy platform that the rest of the site is built on.

Legacy integrations take longer. Your goal with a prototype is to validate if it's worthwhile. This will take more work if your integration exposes bugs or side effects in legacy systems.

New designs get coupled to legacy issues. For instance, you may have a complex registration funnel for your main product. But your new feature may not require that, and might be better off with a simple process. You may find that the simpler registration feeds the more complex multi stage process, rather than pushing full registration for a new feature.

Deep integrations can mask issues with your prototype. You may see user adoption, but only because the new feature is very visible to your installed base. The prototype might be flawed and produce false positive results because of the legacy integration.

Modern platforms with better features are ignored. You may have a poor search function on your site. Requiring the prototype to use the legacy search isn't really necessary. You can take advantage of newer versions of software by limiting their use to your prototype.

You can break legacy inertia with prototypes. If your prototype requires newer versions of platform components, you will build more company expertise in the new version. This will make it more likely that you can migrate legacy code to new versions.

Build new prototypes in isolation, so that you can develop faster, learn more and make better product decisions.

Permalink | Leave a comment »

Learn from success, avoid failure

Mon, 04 Feb 2013 10:40:00 -0800

It has become prevalent to use "failing fast" as jargon for "learning". It is not valuable to fail. Failure means "lack of success". It doesn't mean learning.

When something fails, it is common to identify the failure, and then state that you learned from the failure.

You learn nothing from failure.

Perhaps you have a theory on why you've failed, then make an effort to fix the failure. If you are successful, you will have learned something.

You learned from the success, not the failure. Failure is not sufficient, or even required, for success.

To succeed, prefer to study success. Avoid studying failure, it is a waste of time.

Permalink | Leave a comment »

Why this is short

Sun, 27 Jan 2013 12:53:00 -0800

It's easier to communicate an idea when it is simple and short.

It's ok to write a few sentences, instead of a few paragraphs.

You will choose your words more carefully, and make fewer mistakes.

Permalink | Leave a comment »

Quora Email Settings Anti-Pattern

Wed, 23 Jan 2013 15:51:30 -0800

I can only imagine the regression tests

Permalink | Leave a comment »

"TLDR" is unnecessary

Sun, 13 Jan 2013 14:20:00 -0800

If you summarize your idea in your first sentence, writing "TLDR" is superfluous.

The English Language already has tools to summarize and communicate your thoughts. Leverage existing conventions, which are accepted by a larger population, and you will engage more users and succeed in communicating your idea.

"TLDR" stands for "Too Long, didn't read", and began as a response to the "wall of text" style of writing that would appear on sites like Reddit. Over time, it became a shorthand way of saying "Here's a summary.".

However, if you actually wrote the article, it is asinine to preface it with "Too Long; Didn't Read". Of course you read it. You wrote it.

Instead, edit your article, and make sure to summarize your point in the first sentence.

Permalink | Leave a comment »

How to solve this Facebook Pages error - Our automated system will not allow the name

Mon, 10 Dec 2012 09:53:31 -0800

Recently, I was setting up a new Facebook page with a naming convention that capitalizes various words.

For whatever reason ,Facebook has decreed that this is not the appropriate naming convention.

This is easily worked around.

To fix, accept their selection.

Then go to your page, click "Edit Page->Update Info"

Under name, select the name with casing you wish to have. In the "Edit" window, it will allow you to select whatever casing you want.

Permalink | Leave a comment »

IIS Express does not respect dynamic routes

Sat, 24 Nov 2012 11:04:00 -0800

If you are using IISExpress for your web server during development, and you try to route static files to dynamic files, IISExpress will not respect rerouting those files.

Ie, if you have a dynamic robots.txt file like this:

routes.MapRoute("Robots.txt",

"robots.txt",

new { controller = "Home", action = "Robots" });

You will get a standard 404 error which tries to map it to a static handler.

To fix, you need to configure web.config like so:

Permalink | Leave a comment »

FeatureShot - Web and App Feature Snapshots

Fri, 16 Nov 2012 15:36:00 -0800

I've been working on a new website called FeatureShot.

As I've been building apps this past year, I've been overwhelmed at how fast website and apps are innovating. It is really hard to keep up, let alone discover new things or incorporate new trends into products I'm building.

So, I built a tool that is aimed at developers and product managers who spend a lot of time evaluating new features of websites. It allows you to take screenshots and categorize web and app features.

It is also a useful tool for anyone who is interested in seeing an "inside look" of new apps and website features.

If a product is big and complex enough, it is overwhelming to try and review it. Instead, you want to review a small, digestible part of that feature. I call those reviews "FeatureShots" - they are simply a few screen grabs of a feature, with a bit of explanation around what it does.

My vision is to build a comprehensive feature graph that connects these features. That way, you view all the features that make up a product, compare it to other features in other products, and discover new features that you might like.

Please check it out, and let me know what you think.

Permalink | Leave a comment »

Failcon 2012 Attendee List

Sun, 21 Oct 2012 20:24:39 -0700

I'm going to Failcon 2012 in SF tomorrow and was looking at the eventbrite page and found it was difficult to view all at once.

I created this Google Spreadsheet for the FailCon 2012 Attendee list by scraping that page

Enjoy!

Permalink | Leave a comment »

Stats on announcing CorpQNA to SeattleTechStartups

Thu, 18 Oct 2012 13:18:00 -0700

Last friday, I posted my private beta test of CorpQNA on Seattle Tech Startups list. I've been on this list for many years, and even though I no longer live in Seattle, I feel like it is one of the most supportive and valuable startup resources I've experienced.

A couple of people were interested to know in how this offer was received by the community. I'm just quickly sharing a few stats in case you are thinking of sending an "early beta" product announcement of your own via STS. Hopefully you can learn something.

I feel like real numbers are valuable. Because I am bootstrapping this on my own, I don't have to ask anyone if I can share these numbers. :)

50-100 clicked to CorpQNA's home page. (I don't have exact metrics here, since I didn't put a tracking code on the link)

10 people clicked on the "offer link"

6 signed up for the beta

5 people emailed offline with feedback

0 people accused me of spamming the list. :)

I feel pretty good about these numbers. Most of those people have engaged with the product at least once or twice over the past week.

Overall, I've got about 15-20 companies signed up at this point. Half of those were just people who signed up and haven't come back. The other half are marginally interested. A couple are decently engaged.

Some other learnings I've had:

Spending face to face time with potential users is more valuable than online channels.

If you can pitch someone directly, they will just go to your site and signup. You won't have to worry about a great landing page yet. This especially important when you are still validating your MVP and don't want to spend a lot of time building marketing collateral that may not matter.

Enterpreneurs are generally more supportive

They are willing to try new products, compared to "normal people". I think part of that is because they know how hard it is to get started, part of that is they are very curious about what's new in the market. Though they don't necessarily represent your target market segment, they are really valuable early adopters.

Adwords can be a challenging channel to get beta users.

It costs $15/click in adwords to bid on "collaboration software".

They have a lot of very immediate "bounce quality" metrics and will turn off your keywords if they are not creating a good user experience. I haven't focused too much on that landing page experience and it doesn't perform well on its own yet. If you aren't polished enough, your ad won't perform. And if you spend time polishing your ad too much, you aren't working on your product. If you are entering an established space., you also either have to pick keywords that are highly qualified (which are expensive), cheap (which are not going to convert) , or undiscovered (which means low volume)

The adwords interface can also be a bit overwhelming for a new user. It's a "dashboard" designed for SEM professionals who spend all day managing keyword spend.

Facebook is a little bit easier
Facebook has a "technology early adopter" segment in their ad tool. This is a nice shortcut and feels like it's designed for beta tests. But, If you are going to use Facebook to show ads, don't put an ugly logo up there. Put a picture of a person instead. Make sure you have a call to action. Your CPM rates change dramatically based on clickthrough and this stuff helps.

If there's some stats that you want me to share, or questions/suggestions you've got about CorpQNA, please let me know (via email, twitter, etc)

Permalink | Leave a comment »

22 tools you need for your ASP.NET Startup

Wed, 10 Oct 2012 10:21:00 -0700

My latest company, CorpQNA - Social Q&A For Business, is built (mostly) on the .NET stack. This article explains some of the components I used to pull it together.

ASP.NET MVC

If you are still using webforms, wean yourself away. When we first built PayScale, we used classic ASP, which was closer to the right model than WebForms was. Later, I made the fateful decision to move to ASP.NET Webforms. I was not happy with this decision. Trying to debug ASP.NET Webform Lifecycle events (Page_load vs control_load vs onLoad vs ...) will make your head spin and send you fleeing to Django or Rails.

AppHarbor

Right now, I like it better than Azure Websites and cheaper to start with than AWS Elastic Beanstalk. It's the option that's most like Heroku (which is the gold standard)

Azure Websites doesn't support SSL yet. All of your domains have to be on mysite.azurewebsites.com. Not a good way to project confidence. On the downside, appharbor only supports SNI for SSL ($10 / month) or IP based SSL ($100 / month) Hostname SSL (which Heroku supports for $20 / month) is not yet available. SNI is a problem because older browsers will give warnings when you are trying to use it. I'm using it, but eventually, I'll have to switch. I'm hoping that AppHarbor catches up before then.

AppHarbor is aware of your branching schemes, and you can tell it to listen on a particular branch. So if you are pushing to staging, just push your code to a staging branch and AppHarbor will only deploy changes that occur in that branch. AppHarbor also has a continuous integration server built it. If you use NUnit, this saves you time and helps you avoid breaking your site. AppHarbor seems to use SQL Express as it's database, which seems like it will be less robust in the long term, than SQL Azure. AWS has an RDS implementation for SQL Server, but I think it's the most expensive of the three.

BizSpark

Visual Studio and MS Office for free. The application process is a little bit burdensome but it's worth it. There's also a lot of goodies attached (free credits on Azure, opportunities for networking, promotion and guidance) so it should be a no brainer.

FogBugz

An enterprise class bug tracking tool. I've used FogBugz for 8 years now, and it works very well.

It's also free for "startups", which means for 2 users or less. So there's no reason to not get started on it.

It comes with an integration with Kiln, which is Fog Creek's fork of Mercurial + code reviews. I haven't much used these features, but it's nice to associate my checkins with bugs.

NuGet

Like RubyGems for your ASP.NET application. It's super easy to install new functionality in your application. Install-Package EntityFramework and away you go.

VsVim

If you haven't joined the keyboard revolution (vim, emacs, etc), it's time to get your hands off the mouse! You'll need to spend a few weeks learning how to use Vim (try the great Vimcasts website to get started). Download vim first. Once you are there, look at ViEmu($) or VsVim(github). These are "hybrid" vim solutions. Some commands are still missing and you can't use macros, but the core hjkl type navigation is there.

Airbrake and NewRelic

Stop monitoring your own errors and performance. In the past, we've had to build these tools ourselves (error log monitors) or tried to hook together various open source toolsets (like Ganglia). Do yourself a favor. Until you get an IT professional working full time, use Airbrake and NewRelic. They are free to start and you won't think about that stuff for a long time.

Mailgun or SendGrid

Mailgun can send and receive mail for you. If you are hosted anywhere in a cloud environment (AWS, Azure, AppHarbor, Heroku, etc), using a service like Mailgun or SendGrid is a requirement, because otherwise, your email will get marked as spam. Simple as that. You can also you webhooks to let your users use email as an interface into your product (instead of having to visit a website). They also even parse out the junk from your email, so you don't have to. (signatures, Subject lines, etc)

Ninject

Ninject is Dependency Injection made easy. There are a lot of frameworks out there for dependency injection. Fortunately, they all require your code to look almost identical (lots of interfaces, avoiding instantiating objects directly). So which ever one you use, it's not that hard to swap out for another one. Ninject is super easy to get started with. So if you are no to IoC or DI, it's a good place to start.

Also, If you are planning on doing any sort of testing, you'll need a Dependency Injection Framework. Otherwise, your test setup scripts will start to look like a mess.

NSpec

Still using NUnit for testing? Time to take a look at NSpec. This is based on the hugely popular RSpec library for Rails. Your tests will be a lot cleaner with NSpec. You can set up elaborate contexts to reproduce difficult errors. It's not as supported as NUnit, and there are other BDD Frameworks available which might be better.

Intercom.io

I was looking for a customer feedback tool, and was originally planning on going with UserVoice or GetSatisfaction. But then my friend Lee mentioned Intercom. This is by far the best way to get feedback and communicate 1-1 with your users. It's not cheap but if you want a good relationship with your user base, this tool is a must have.

Twitter Bootstrap

The pre-eminent framework has lots of best practices built in. In addition to making your website easier to read and navigate, you'll learn a lot about usability and design simply by implementing the framework. There's also a growing ecosystem of tools, themes, and plug-ins that work with Bootstrap. Bootstrap has also recently moved into it's own foundation, so it should be around for a while.

Entity Framework or NHibernate

If you are developing a new MVC application, you are either going to choose Entity Framework or NHibernate. Frankly, for a long time, NHibernate was the better choice. It probably still is if you have classic N-Tier SOA application. It just works better with disconnected datasets. It has better support for caching semantics and more fine grained control over your apps. But it can't be a bear to get started with.

If you are just starting out, Entity Framework with Code First models will get you a long way quickly. The LINQ oriented syntax is clean and easy to understand. But there are still the occasional side effects related to how Entity Framework stores and manages objects under the hood. Hopefully this stuff will get resolved. The EF team seems to iterate very quickly and is able to get language-specific features into the .NET platform.

The other caveat is it is really hard to switch from one to the other. if you use a Repository pattern, this can help. But a repository pattern doesn't really work well with either ORM, and requires a LOT more code. So just plan on sticking with one or the other for a long time.

Postal or MVC Mailer

Rails has ActionMailer, which lets you write emails using typical MVC semantics. Unfortunately, this is not yet built into ASP.NET MVC. But there are some good libraries to handle it. I use Postal, which leverages the open source RazorEngine library to create your own Razor views. The syntax is slightly less powerful than Razor for ASP.NET (ie, no Html helpers, layouts, etc) but it's better than concatenating strings together to produce your emails.

Glimpse

Like Firebug or Chrome Dev Tools, but for your ASP.NET website. This is a great way to visualize your performance.

SOLR and SOLR.NET

Microsoft doesn't offer a world class search engine. SOLR has all the features you'll need. SOLR.NET is a wrapper around SOLR and makes it easy to integrate. Not as easy as Sunspot for Rails, but you can get going with SOLR in less than a day if you code well.

Linode

For $20 a month, you'll get your own linux server. Hard to beat. I use it for lots of utility functions, as well as hosting my own SOLR server. There are a ton of "cron jobs" masquerading as subscription cloud services. If you don't watch out, you can get dinged by 5 or 10 services that each cost $10 a month and all of the sudden, your hosting bill is through the roof. BTW - oddly enough, Azure is offering a linux VM even cheaper than linode, at $10-$15 a month. But it's only available in preview right now.

Did I miss anything? Continue this conversation on Hacker News.

Like this? For more, follow me on Twitter @tarr11 or on Google+

Permalink | Leave a comment »

Is Google Talk a LinkedIn killer?

Fri, 05 Oct 2012 11:01:00 -0700

Google Talk has stealthily become a ubiquitous professional networking and business communications tool.

If I've ever had a two-way email conversation with someone in GMail, they show up in my GTalk list. I also see people who are in my G+ circles.

Since then, my Google Talk contacts list has been growing, and is more of a recent representation (but incomplete) representations of my connections than either LinkedIn or Facebook.

Some of these people I've never met. I've talked to some, once or twice about various business ideas. Some are parents of kids on my soccer team. Others appear because we replied to some threads on various discussion lists.

Unlike LinkedIn or Facebook, I didn't have to make a decision to formally "connect" with them in order to maintain a relationship. I just had to send them an email, and have them reply.

I've added a few new contacts on LinkedIn (almost none on Facebook), but there are always new contacts appearing on my GTalk list.

I get to know someone a bit more personally when they are on a chat list. I see when they are online and off, their G+ profile, their GTalk status update. It's also integrated with email, which is my primary communications tool, which makes it frictionless to send them an IM, email or video chat.

It's become more important to me to get people on my GTalk list, rather than on my LinkedIn profile.

How are you using Google Talk for professional networking?

Continue this conversation on Hacker News

Permalink | Leave a comment »

undo.io is now on github

Tue, 15 May 2012 15:02:00 -0700

I've open sourced undo.io. Here's the github repo.

A place for my text files

Undo was just going to be a place on the web for my todo.txt files. It still is that, but I added a bunch of features, some good, and some, well, less good. I wanted something that let me mostly work in text. I am a decent typer, I really like playing with new text editors (hello Sublime Text and Notational Velocity!) I also felt that text was future-proof, so there'd be less concern about what happens to your data if the service goes offline.

People often ask me why i called it undo. I wanted to find a way "undo" my todo lists, rethink them, and make them simpler.

A way to learn Rails

I've been working on undo.io for about 4 months and it was my first Rails app. I learned a lot (Rails, Linux, Vim, Heroku, etc) and i think I'm a better developer for that.

A hobby is not a business

I learned was that undo is not a business - it is a hobby and part of my workflow. I still think productivity is an interesting space and notes in particular are so embryonic, there's always some interest. That's the challenge - I didn't really publicize the tool other than a link or two here and there, and still, people come. I think if you write a note taking app, a core group of passionate people will try it out, and some may even use it long term, just because that's how this product space is. Productivity is a very personal thing - everyone has a different idea on what is right and it's fun to learn about. I still think there's some value to the product and some interesting ideas. I still need a note taking app and this will probably be part of my workflow for some time. I'm putting it up on github since I think it could use some daylight. I also feel like github is my new resume so I'd like my work product to be up there. This isn't my cleanest code but when working as a lean startup, that's probably a good thing.

Existing users

Note to existing users: I'm not currently planning on taking it down for the time being. I'd suggest you sync your notes to Dropbox though. That way, you'll always have a copy of your data.

Moving!

BTW - I am going to be moving my family back to the bay area in July. I'll be really interested in meeting with people to talk about new opportunities and interests. Stuff I'm interested in working on / learning about: big data, hadoop, bayesian networks, statistic programming, NodeJS and js projects in general, mobile development, HTML5 / CSS3, NoSQL.

If any of this sounds relevant or you want to get in touch, please ping me

Permalink | Leave a comment »

Local First

Wed, 02 May 2012 10:43:00 -0700

Local First

With the advent of HTML5, localStorage, mature javascript client libraries, ubiquitous REST APIs and mobile apps, we don't need to rely as heavily on cloud providers to develop our applications It's time to change how we are writing our apps.

What is local first?

Similar to mobile-first (where you build your app to work well first on mobile, and then expand out to other devices), Local First is the idea that you can build a web app that works first as a disconnected app. You can synchronize to cloud providers if you want, but your app should work locally.

What are the concepts behind local first?

Cloud-optional

- Your app should still work without cloud storage.

Data-permissions

- User data should not be sent to a cloud provider without permission.

Network-optional

- Once installed, apps should run without need for an internet connection.

Cloud-choice

- Users should be able to choose appropriate cloud providers for data storage.

Cloud-provenance

- Cloud providers should clearly document where your data is stored in understandable language

Why Local First?

There are many reasons to build your apps local first. Performance and Stability, User Adoption, and Privacy

Performance and Stability

Local apps will have less dependency on network calls. This will (usually) make them perform better. If your network goes down, or is slow, firewalled, or otherwise encumbered, your app still works. Less headaches for you as a developer.

Adoption

Users are increasingly concerned with their privacy. Your application is probably losing users since they are uncomfortable with how you are using their data. Many corporate networks simply will not allow your data to be stored on cloud providers (other than their own networks).

Government privacy concerns

Government snooping continues to increase. It's a losing battle (PIPA, CISPA) to fight governments who are intent on monitoring our internet usage and data. As more governments gain control of the internet (with different points of view on the importance of free speech and privacy), your data will be in more hands. By all means, let's continue to slow them down. But more importantly, let's make sure our data is not centralized.

What's next?

The idea here is that this is not a centralized approach, so the conversation shouldn't be centralized as well. As a developer, the next thing is to try and build an app this way. Don't start with a SQL database and server application. Start with a client application that works without the cloud. If you've already been doing this, share how you do it. Post your code. Use #localfirst in your tweets, or somewhere in your posts, so we can track local first activity.

What I'm going to do next

I've built my latest app undo.io in rails as a web-first app. I'm going to spend a few weeks flipping it to local first. I'll share how that works on my blog

Discuss on Hacker News

Permalink | Leave a comment »

We tweet in code

Wed, 25 Apr 2012 14:22:00 -0700

Are you a developer or tech type? You probably have a twitter account (here's mine). You might even have a LinkedIn account.

Do you have a github account?

Ever since I've started to work on undo.io, I've started attending lots of events and taking a fresh look at the job market. I noticed the trend a few months ago after attending some Seattle Rails and NodeJS events. As I'm relatively new to the openstack community (coming from windows/.NET) I've been asking for quite a bit of help (people have been really awesome by the way when you ask). The most helpful things I've found are where someone points me to a relevant repo on github.

Github has infected many conversations I have with developers now. Developers don't say "follow me on twitter" or "check out my linkedin account". Instead, they'd point you to their github account, so you could see their latest nodejs project, .vimrc or DCPU-16 hack.

Code is how developers communicate, and feel most comfortable. Compared to how much we code, we don't really email, tweet, post status updates, or blog posts. README files and comments in code are where we make our jokes, and we define ourselves in terms of languages we know, operating systems, or text editors. (Don't believe me? ask a dev about any one of those topics, see if you can get them to shut up)

There are an increasing number of forms that have asked for my github account as a supplement to a traditional resume or linkedin profile. It's a really great way to find out who you are as a developer, and what you are interested in. Even if you aren't committing to a major project, looking at what you are watching, in which areas, is incredibly valuable.

As a hiring manager, it lessens the need for code samples or programming interviews. After all, there's your code, right up there on github.

Having an active github account doesn't mean you'll get the next job, contract or round of funding.

But if you don't have one? You might get disqualified altogether without even knowing it.

Read my notes for this blog entry on undo.io

Follow me on github

Discuss on Hacker News

Permalink | Leave a comment »

We miss you! NOW BUY SOMETHING.

Fri, 02 Dec 2011 11:45:00 -0800

Here's a stream of messages from various companies telling me that they miss me. I never call, never write, and most importantly, I don't follow their calls to action. Alas, email marketer, you don't miss me. You don't even know me. So quit it with the guilt trips.

Permalink | Leave a comment »

Boardroom in my brain

Fri, 11 Nov 2011 10:19:37 -0800

Well now I have to see if I have the nerve to stick a knife in my own idea. Again.

I hit these moments. I already killed a few ideas. Most of those never made it past Balsamiq. But this one is already up on staging in Azure, costing 12 cents an hour + a small SQL instance. I've got 6 different csproj files and even some unit tests.

I'm in the middle of some coding sprint and some roadblock makes me look up from VS2010 and question the whole thing. Someone built the same thing I was thinking about. But they did it 4 years ago. I am more behind the curve than I thought. 38 years old in a few days - what are the 21 year olds working on? Would I even understand? Can I pivot from 2008 to 2012?

I envision a room full of the smartest people I know. There is a powerpoint projecting on a wall, maybe a plate of cookies in the center of the room. I recreate the whole boardroom in my brain, because right now, I am my own board. Someone has to call me out on my BS. Someone who hasn't spent nights and weekends coding away and fixing a million bugs and learning SOLR or Entity Framework just to get this thing off the ground.

They will ask me things like - How is your product differentiated in the market? Have you looked at [X]? Seems just like yours! How are you going to acquire customers? How long is it going to take you to get to market? Or worse yet, they sit there, tapping away on their iPads, totally and completely bored.

I used to leave these most of these questions to someone else.

Oh yeah, that's me now, isn't it?

Permalink | Leave a comment »