I’ve been on the scout for a nice way to capture, prioritise, schedule and track tasks. My requirements were fairly simple –

• Be able to capture tasks quickly
• Be able to prioritise tasks
• Be able to schedule tasks
• Be able to track tasks
• Must integrate with my Blackberry.

I had a look at a few solutions, but the one that really impressed
me was Remember The Milk. Let’s skip past the hyperbole and dig
straight down to why I like it.

Remember
The Milk (RTM) allows you to create as many lists as you like. I have a
quite minimalist approach to lists, but in your Task view, they are
simply tabs:

If tags are more your thing, you can add any number of tags to a task, and there’s a helpful tag cloud for familiar navigation.

Task details are easily set:

RTM integrates with GMail (including Google Apps). This wasn’t a
requirement by any means, but considering that it’s one of my most
visited web pages, it was a most welcome addition.

Tasks are easily prioritised and colour coded.

I
would heartily recommend that you spend a few minutes to become
familiar with the keyboard shortcuts, they will dramatically improve
the speed with which you can use the site.

The RTM web interface
is good and I use it every day, but it’s strength is in its
integration. It integrates with my Blackberry (yay!), Windows Mobile, GMail, Google calendar, iGoogle, iPhone/iPod Touch, standard mobiles and Twitter. You can receive reminders via email, SMS, and instant messenger (AIM,
Gadu-Gadu, Google Talk, ICQ, Jabber, MSN, Skype and Yahoo! are all
supported). In other words, you would be hard pressed to not be on top of your task list.

Some
of the integrated services are available only to pro account holders
(of which I am one) – but considering the price tag – $25/year – I
would say benefit to cost ratio is sky high, so why not sign up for a free account and give it a whirl.

What do you use for task management?

Image via Wikipedia

The third and final part of this mini series (trilogy?) is on
reverse tunnelling, or remote forwarding. Let me give you an example of
when this might be useful:

Let’s say you’re on a client’s site. Let’s say your client has a
build server that’s only accessible from the internal network. Let’s
also say that your client is a humungous corporate blob, with 8
signatures and 6 weeks lead time in getting VPN
set up for you. Or maybe if you’re a consultant they won’t let you have
VPN at all. Mentioning no names…. – You need to be able to kick
off a build at some point over a weekend. And since the client’s site
is a gazillion miles from your home, you’d rather not hang around.

No problem. SSH is here to save the day. Our hypothetical build
server sits on 192.168.1.123 on the client’s internal network. It is
also set up for access via Remote Desktop. While you’re still in the office, we establish a session thus:

putty username@mydomain.com -R 9999:192.168.1.123:3389

This is telling the sshd daemon on your home server to bind to port
9999 and listen for incoming connections. Any incoming connections are
forwarded through the tunnel, to port 3389 on 192.168.1.123 on your clients network.

My home server runs at 192.168.1.101 on my network. So when I’m at
home, if I want to connect to my client’s build machine, I instruct my
remote desktop to connect to 192.168.1.101:9999 and as if by magic, I’m
now on the build machine. Awesome.

SSH is a really awesome workhorse, and if you’re a web worker,
consultant, micro ISV etc I really recommend spending a few moments to
get to grips with it.

Image via Wikipedia

OK, so last time we looked at how to tunnel your web traffic over
SSH to increase security, privacy, network restrictions etc. Today,
we’ll look at how to use port forwarding to access servers that are ordinarily available on the internal network.

Subversion Over SSH
As an example, I’ll show you how I access my internal network’s Subversion server. Here’s my set up:

• Externally, I have MyDomain.com which resolves to my home IP.
• On the internal network, my Ubuntu server has the address 192.168.1.101.
• The server runs an OpenSSH server and the Subversion svnserve daemon.
• The router forwards incoming connections on port 22 on to my server.

Setting up the tunnel, we do the following:

putty -L 3690:localhost:3690 username@MyDomain.com

Log in as usual. This time, the SSH tunnel
has been set up such that any connections to 127.0.0.1 on port 3690
will be tunnelled to port 3690 on the server. Which, handily, is the
default port used by the svnserve daemon.

And to get hold of some part of the remote repository, we just do:

svn co svn://127.0.0.1:3690/MyRepo

And hey presto, tunnelled access to your internal SVN server. It’s
always puzzled me why the good SVN folks suggest their weird svn+ssh
hack, which launches an svnserve instance in user space, potentially
causing file locking problems and requires user access to the svn
database files.

This technique can be applied to any internal server, whether web, database, remote desktop etc.

Next time we’ll look at reverse tunnels. Very cool.

We’re going to need 2 things. A server running OpenSSH, and an SSH client. I have an Ubuntu server under my stairs, and I have full shell access at my shared server at DreamHost,
so I have 2 options, depending on what I’m doing. We also need a
client. If you’re running Windows, you will need to grab Putty. If
you’re on Linux, you’re already set. I don’t use Macs, but I’d be
surprised if they didn’t also ship with an SSH client.

Tunnelling Your Web Traffic
Maybe
you’re in Starbucks using an unsecured wifi hotspot. Maybe you’re
behind a fascist corporate proxy. Maybe the NSA is watching you.
Whatever the reason, tunnelling your web traffic with SSH is trivial.

You
see, every SSH client has the capacity to turn itself into a SOCKS5
proxy. Nifty eh? What does that mean? Well, if we launch the client
like this:

putty -D 8000 username@myserver.com

You’ll be
prompted for your password, just log in as normal. It just looks like
you’re into your shell. Something else has happened though – your
client has set up a SOCKS5 proxy accepting connections on 127.0.0.1
port 8000. Any applications set up to use this proxy will send their
traffic through the SSH tunnel, and out the other side at the server end. Let’s try it.

Go to www.whatismyip.com. This is your IP address. Now go to wherever in your browser you set up a proxy, and enter 127.0.0.1:8000 under the SOCKS proxy entry. Here’s an example of doing it in Firefox:

Now reload the page at www.whatismyip.com. The IP address it reports
should have changed to the IP address of the SSH server. Your traffic
is now tunnelled.

Is that it? Probably. The only limitation of this is that DNS traffic is not tunnelled. This is what is usually termed DNS leakage.

But Firefox, bless its cotton socks, has a trick up its sleeve. If
you type this in your Firefox address bar, and click past the amusing
warnings:

about:config

You’re into the Firefox configuration area. You could really bust stuff in here so pay attention

Scroll down until you see this:

network.proxy.socks_remote_dns

Change the value to true. And that’s it. Firefox is now kindly
tunnelling your DNS traffic through your tunnel, and you’re super
secure. Even a rogue hotspot with poisoned DNS can’t hurt you know. Cue
evil laughter.

Next time I’ll show you how to turn SSH into a poor man’s VPN, but without the hassle of setting up a VPN server.

Image via Wikipedia

It’s been a busy old week. I’ve been applying for my Authorize.net
gateway and merchant account, e-Junkie, arguing with PayPal, writing
the C2X website, and many other wonderful things. Over the last month,
I’ve discovered 3 things I didn’t know I couldn’t live without.

DreamHost
I have thus far
been very happy with DreamHost, to the point where I have moved all my
hosting (with the exception of this blog) to them. The support is
quick, the one-click installers work a treat. I particularly like the WordPress and Subversion
installers. Now, I host each project’s Subversion repository under its
own domain at DreamHost. It’s fantastic. The full shell access is also particularly useful, and uncommon for a shared host.
Top marks guys.

VisualSVN
I’ve been a happy AnkhSVN user for years. AnkhSVN, for those who don’t know, is the Open Source Subversion plugin for Visual Studio.
The writers seem, for whatever reason to be struggling with VS2008
support. I persisted with it as long as I could, but it trashed my
working copy too many times, so I had to ditch it. My colleague Rob
introduced me to VisualSVN, a very reasonably priced plugin for visual
studio that integrates the TortoiseSVN
shell extensions into Visual Studio. Works like a dream. The folks
behind it seem pretty switched on too. Their company is registered in
the BVI, which tells me they’re not afraid to think differently, and
their customer service is good. I emailed them and said I wanted to buy
it, but didn’t want to pay through Digital River for fear of being scammed, so they emailed me a link to pay through Plimus. Nice work guys.

ReSharper
I’ve
been developing commercial .Net applications since 2001. That’s 7 years
of working without ReSharper that I can’t get back. People have been
telling me for years to use it, but for some reason, I didn’t. Thanks
to Rob again for convincing me. It is truly awesome. Awesome in a way
that would make you a fool for ignoring it for 7 days, let alone 7
years. If you don’t have it, get it. Really. No, REALLY.

Image via Wikipedia

I’m trying to think, I think I’ve been using GMail since 2003. It’s been a great service and I’ve loved it I use Google Apps for all my domains, and have found them to be also flawless. I know Google have been having a few issues, but I hadn’t seen any of them. Until tonight.

I’m trying to communicate with Authorize.net about my merchant account
application, but with the email down, I’m cut off. It’s frustrating,
and it’s hard to not feel worried about this spate of failures in so
short a time frame, given that my business may soon depend on email.

Image via Wikipedia

A while ago I gave someone some advice on how to get fit if you do
an IT job. I said that for me, saying “I’m going to get fit” doesn’t
work, it’s too fuzzy, too wishy washy, there’s no time frame, no sense
of success or failure. What worked for me in the past is this:

• Find a sprint triathlon in your area that’s about 2 to 3 months away.
• Enter it. Actually send off a cheque – commit.
• Tell your friends and family you’ve entered it.
• Tell your blog readers you’ve entered it.
• Go to BeginnerTriathlete.com and get a couch-to-sprint training program.
• Stick to the training program religiously.

I found that the fixed time frame, combined with the knowledge that
friends and coworkers were going to check my performance on the web
site after the event was a huge driver. The fear of failing publicly
was too much to bear. I trained. My God I trained. The event nearly
killed me, but it worked. It absolutely worked.

I’m finding a bit of a similar feeling creeping in on what I think of as my micro ISV project. A little sensation of drift, a feeling that there’s no time frame involved.

Enough. Mark my words well. Failure is not an option. It’s time to get serious. It’s time to get specific.

Home Document Manager v1.0 will be released on the 1st of August 2008

I feel  a little bit inspired by the recent 30 day sprints. I’m cheating a little, I have 34 and a bit days to do it.

34 days and counting.

Whilst I’m still doing contract development, I find that it really helps me look at the true cost of my activities to my business.
For example, were I to code the whole product myself, I would be coding
Windows  shell namespace extensions by hand, writing a Winsock
layered service provider as well as having to implement  a SSH
client from scratch. 

The NSEs are doable by hand, but a royal pain in the behind when
you’re using managed code, and can be fiddly, error prone and time
consuming. Writing an SSH client from scratch would require a lot of
research and development time. Writing Winsock LSPs is such a tricky
niche area that even the Microsoft samples that ship with the SDK are
buggy. My experience as a contract software developer has made it
easier to put a £ (or $) sign on my time. Outside London, an average C#
development contract brings in about £500 per day. So that’s how I value my time.

An expert is a person who has made all the mistakes that can be made in a very narrow field.

Niels Bohr (1885 – 1962)

Buying in a namespace extension component has set me back about half
a day’s earnings, the SSH component about the same. The LSP is the big
one as I’m having it written by an expert, this will cost about a day and a bit of contracting time. So less than 3 days of my time has paid for all the tricky bits of my product,
leaving me to concentrate on the plumbing. It would probably have taken
months for me to learn and perfect these individual components if
written in house. Now of course, there are all the questions about
ownership, bug fixing, enhancements etc, but I need to get to market
first, and I’m going to take the quickest route possible to get to 1.0.

What experiences have you had with outsourcing bits of your micro ISV product?

Typically, we write down user stories on index cards – these cards
will include a brief synopsis of the story, and an estimate as to how
long it might take.

Some practitioners prefer real time estimates, some prefer a points
system. I fall into the latter camp. For any story, I can award it a
points value of 1,2,3,5,8 or 10 points. 1 Being a trivial story.

10 is the interesting one from the point of view of this post. 10 means it’s an epic.
It serves as a cue for the devlopers to request that we talk about the
story, and break it down into smaller stories, so that they can have
meaningful estimates applied to them.

A post on Web Worker Daily
got me thinking about how epics relate to the dreaded task list. What
the points scale to in real time depends on the team, the iteration
size, the team’s velocity etc. In a Micro ISV, we’re spread pretty
thinly, so I would think we should really say any task that would take
5 days or more to complete should be classed as an epic. When time is
tight, we should look for demonstrable progress in short time frames,
so tasks should perhaps be more granular than in an enterprise
environment.

Teams move boulders. The micro ISV fills his pockets with pebbles and runs back and forth.

“Your touch typing is better than mine” I quipped.

“Fifty years o’ practice” she replied.

During my 2 hour wait to see whatever 17 year old they just made a
doctor, it occurred to me that of the large number of programmers I
know, comparatively few are touch typists. It’s very strange how this,
the most fundamental of skills, seems to be neglected in favour of the
latest Ajax book, or Petzold offering. I had a look around to see what
resources I could find to help improve my typing skills.

There were a number of commercial offerings, but for some simple drills in a nice web app, I’d recommend this:

http://www.sense-lang.org/typing/