Note to self 15 - If a project has no risk - don't do it.

Dear Michael,
Liz Keogh in her QCon London session on different kinds of incentives last week reminded me of this quote from Tom DeMarco and Timothy Lister's "Waltzing with bears".
As DeMarco and Lister put it in the book each and every project that is worth it's while has to have risk.
For me the implication of a project with "no risk" is not only, that it could be done by anyone in any environment but also that it doesn't matter if this project fails - essentially that is what "no risk" means in the end, right? - and who would want to work on such a project?

Digital Taskboards I'm aware of as of 2011-03-11

At the Agile London usergroup meetup last night the topic of digital taskboards came up and I promised one of the participants to send a link to /one/ of them… since I couldn't remember which one I was referring to (it was kanbanery, I found my note eventually) I went through the list in my head and realised that the list isn't so short any more…

• Qanban…a very promising (and free as in speech) project from @xlson which unfortunately doesn't undergo much development right now. But… it's open source, the source is on github - you know what to do!


• Kanbanery…got a lot of mentions lately on twitter, but I only did a testdrive


• Pivotal tracker…A friend of mine uses that a lot (Cheers @jcfischer) and is quite pleased with it


• See Now Do… haven't tried it, but I know the product owner and from that I infer that it ought to be good


• Atlassian / Jira / Greenhopper
… that was the one we where searching alternatives for...

@Lynne: Sorry - to many to tweet directly

Note to self 14 - Backups don't matter, what you want is recovery

It really doesn't matter how good or how fast your backup process is - it's fastest if you just copy to /dev/null - if the recovery doesn't work. 

This is one of the times when the practices of ITIL and the like are very applicable even to the smallest IT department - like your own personal laptop computer.

Do a trial run of the recovery every now and then!

(At least I was told that ITIL emphasizes these trial runs. If it doesn't – well, to do them still seems like a good idea to me)

Edit: As Robby Raschke said on Twitter: IT Service Continuity Management (ITSCM) in ITIL is where testing your business recovery is a critical component.

Note to self 13 - you can't assign responsibility, it has to be assumed

Michael, you might be able to say - or even order - that somebody is responsible for something but that doesn't really make them feel responsible for it. If he is assigned that responsibility without assuming  the responsibility he will most likely find ways to delegate this responsibility and to redefine the borders of the responsibility. 

An example for assumed responsibility would be this guy I know who develops his own software for a living - they lengths he goes to circumvent the glitches of the operating systems are impressive. For example even if the java installations on the client machine are completely effed his software will still try to figure out the best match for his needs and will 'just work' - Although he could just lay back and take a position of "Well, if the user can't correctly install java it's his problem" he doesn't. He feels not only responsible for implementing his software. He feels responsible for the users experience with his software. (And given the fact that he is able to make a living from a $30 software package his commitment seems to pay off)

It's unrealistic to expect this kind of behavior from people who are assigned responsibilities - especially if they don't believe in the goals or the underlying beliefs of the people who try to assign the responsibilities. 

Big scale 'Agile', enterprise 'Agile' - what went wrong?

I'm just trying to understand the idea of large scale 'Agile'
"Apply the principles that we found out to work so well in small groups - which is what makes small groups so successful, because only in small groups these principles can work - to large projects?"
Seriously?
Whom are you trying to be kidding?

Let me put it another way:
1995 to 2008: 7 plus/minus 2 is the ideal team size because of [the usual arguments]
2008 ff. : 'Agile' works so well, lets just apply it to the enterprise level


====8<--------------
Yes I know: the 7±2 team size thing is way older than 1995 - I was referring to usage in the 'Agile' (capital A) community

Note to self 012 - Write it down immediately - with a pen!

Dear Michael,

your short term memory is mediocre at best - deal with it.
If you've got a bright idea while you're on the phone, write it down. Immediately.
If you've got an important thought while browsing the web, write it down. Immediately.
If you've got a sudden intuition while you're at the beach, write it down. Immediately.

There is a clever toolset designed for this task - it's called PenAndPaper. No, really using pen and paper has some huge advantages over iPhones, Blackberries, iPads and the like.
- You don't have to leave the application you're working with right now.
- You don't have to search for the application
- You can use it while you hold the phone to your ear.

Oh, and it doesn't run out of batteries and even survives coffee spills (albeit at the loss of some legibility if you keep on insisting on using these antique fountain pens)

Therefore, dear Michael: Get back in the habit of almost always carrying a minimum amount of paper and a pen.

P.S.: But of course putting it down on paper doesn't eliminate the need to consolidate the notes at some place - in your case I wuld recommend something electronic, virtual and distributable. Like a plain textfile. A text based mind map might also be in order as long as it really helps and doesn't come in the way.

P.P.S: Oh, and since you're such a keyboard fan - as long as you're on a 'real' computer and can acces your notes faster via a the keyboard that by reaching for pen and paper (e.g. <Cmd>-<Space>, "Scribbles", <Enter> or <Windows>-<R>, "Scribbles.txt", <Enter>) I guess it's OK to use that instead of paper...

Context does matter in UX und UI design

In a recent tweet a friend of mine (Selena Delesie) stated that she is "Surprised to discover there are programmers who still don't put limits, data type restrictions, and error handling on form fields."

As much as I like twitter I soon came to realize that I couldn't put my answer in 140 characters - not even with unicode tricks.
But I'm not surprised at all - and I don't even think it's a bad thing that there are unchecked fields on forms.

As I said in my reply:
"depends on context & is a tradeoff between effort & possible harm.
user ∈ inhouse dev ⇢ lax checks"
user ∈ public ⇢ strict checks [didn't fit in the 140characters limit]

The scenario for user ∈   inhouse dev ⇢ lax checks

This would probably also be a 'C' or a very small 'D' on the cockburn scale, wich is explained in more detail at Alistair's own site.
If I build a 3-hour effort tool for me and my fellow developers to manipulate database metadata and that is intended to run only on our development machines I probably

• won't put too much effort into checks against SQL-Injection - we can do all the harm we want anyway [and it doesn't matter much if we accidentally write "drop database;" in the database's command window or in an unchecked entry field]


• won't put to much effort into checks against the wrong type - I'm pretty sure all of the intended users (e.g. me and three co-workers) can handle an error message like "invalid type at line 8745 in <name_of__3rd_party_sql_library>"


• won't care if it's possible to enter 4gb of data through one of the entry field - the possible harm is well within acceptable limits for the stakeholders (me and my colleagues) and the harm could be achieved in much simpler ways

The scenario for user ∈   public ⇢ strict checks

On the other end of the spectrum would be a piece of software used by a large number of people who might or might not have have malicious intentions and where incorrect values might cause harm to serious money or even life (that would be a bigger 'D' an 'E' or an 'L' on the cockburn scale IMO.
The classic 'E' example for me is the ATM where I probably

• would make sure that only numbers are entered by using a hardware keyboard that only consists of numbers [and check the input values to be a little bit safer with respect to physical attacks]


• would replace drop-down boxes by large hardware buttons at the edge of the screen


• and so forth

The 'L' examples that come to my mind are mostly related to heavy machinery or medical equipment - in both cases the aforementioned principle - limit the choices and represent data entry through manipulation of physical objects - are heavily applied.
Most application fall somewhere in between and nowadays more and more applications that deal with essential money (e.g. online banking) are realized without the hardware representation. In these cases, especially for applications running on the internet, I would

• Put limits on the length of input data - on the client if possible and on the server just to make sure


• Make sure the data that get's typed in represents the correct types * as early as sensible [that might be on the client - I'd add a server side check for good measure in all cases and might drop the client side check if it becomes to intrusive for the user.


• Implement an error handling scheme that gives helpful information to the use with as much detail as he requests and informs the developers and maintainers of the system of the possible quirks in the system at the same time so that the UX can be optimized to have less occurrences of that specific error in future releases.

The problem with the two scenarios "user ∈   inhouse dev ⇢ lax checks" and "user ∈ public ⇢ strict checks"

is that most applications are somewhere between them. And sometimes application evolve from a little developer-on-the-team-only helper application to something that is used by more and more people so the last responsible moment ** to rework my three our project to a product becomes hard to determine. But if I as a developer kept my options open (e.g. because clean coding has become a second nature) it should be possible to productize the tool.
But - it has to be a conscious decision!
In my opinion it neither makes sense to make a cheap tool cat-proof nor is it responsible craftsmanship to offer a product to the public that doesn't employ proper checks.

* but that is a topic for another article - the only type that a user with a keyboard can key in is "array of characters" - what it represents is 100% context: e.g. "feed" might be a word in one case - it's the hexadecimal representation of 65261 in others.

** See e.g. Mary and Tom Popendiek on the topic of the last responsible moment and the keeping of options. Due to a fancy website design on their side I can't provide a link though - but as of 2011-01-22 the statements can be found after clicking on "focus on learning" in the principles box on the left hand side of their website.

BTW: I really wonder how this page with the "is element of" and the arrows will display on other browsers and systems...

Note to self 011 - Trust takes a lifetime to build - but can be lost at the blink of an eye

And - as Jerry Weinberg said - no one tells you when they have lost the trust in you!

Note to self 010 - Parkinson's law strikes - again!

It may not sound very nice, but the sad fact is that people tend to use up the available time for a task in a lot of situations. Therefore: If you want to be over with the meeting in 30 minutes send out an invitation for only 30 minutes - if you make it 45 minutes in the invitation it will take 45 minutes.

Parkinson's original quote from 1955 is:

Work expands so as to fill the time available for its completion.

see: http://en.wikipedia.org/wiki/Parkinson%27s_Law

Note to self 009 - it's called Hanlon's razor

And it goes: Never attribute to malice that which is adequately explained by stupidity.

Although sometimes Heinlein's Razor or Napoleon's unnamed version of it seem more appropriate...

Heinlein's razor:
Never ascribe to malice that which is adequately explained by incompetence, but don't rule out malice

Napoleon's unnamed epigram:
Never ascribe to malice that which is adequately explained by incompetence.

See http://en.wikipedia.org/wiki/Hanlon's_razor

Note to self 008 - It's not "You are " it is "I observe "

Dear Michael, [talking to myself as usual] I noticed that you complain about incompetent people. When I hear that I tend to put you in the category "egomaniac". What can we do about that?

Sounds much nicer than: "Michael, you are an egomaniac. I know that because you complain about incompetent people" doesn't it?

Even if you're convinced that the other Person (or team for that matter) is malicious, unhelpfull or in any other way "bad" remember that that's just the way you perceive it. Talk about yourself - that is something you can be sure about and that can't be argued away. [It is very hard to tell somebody "no, you do not feel that way" - although sometimes there are people who try.]

Note to self 007 - I still can't handle incompetent people

More a point I still have to work on than an actionable note to myself:

Don't get overconfident. It's not enough that somebody is incompetent - you've got to be able to convince people of this seemingly obvious fact. If you can't pinpoint it - just let it go for the moment. Especially if you can't put a number to the incompetence.

As a matter of fact this is something that I've never really gotten around to. 

[Explanation of the note for reader who are not myself] 

There is a certain kind of incompetent people (e.g. they just have no expertise in the very thing they are doing, they are literally incompetent in such a way that they lack the competences that they need to perform their job) that I still don't know how to handle - those that don't know the details of the job their doing but are yet well versed in easily graspable abstractions of the work they should be doing. 

The only way to handle these people - who sometimes have a real bad effect on the morale of whole teams - seems to be a very detailed breakdown of their shortcoming. But that on the other hand leaves me in the position of a nitpicker [erbsenzähler] who wastes everybodys time - including my own - by going through seemingly unimportant details.

Of course - applying the rule of three - that can't be the only solution - but although I've been trying for years I haven't  been able to find any other angles. At leas not up until now - still trying!

Notes to self 006 - Be aware of the focus of the reciepient

In an oversimplified view there are generalists (often sporting an NT of NF in MBTI speak) and there are "detailists" (often sporting an SP or SJ in MBTI speak).
You (I'm talking to myself here) are a NF generalist. If you want to convince an detailist you need to use arguments with lots of actionable detail - not the abstract mumbo jumbo you're used to.

(Wish I had thought about that two weeks ago)

Notes to self - 005, Get it in writing - even if you have to put it there yourself

As I wrote a while back it is always a good idea to get important agreements in writing - Today I just had to remind myself that - especially in these days of electronic communication - there is always the option to put it writing yourself.
As a last resort - e.g. if an official protocol would be taken as an offense - I take the tangential approach as in "When we agreed on we also mentioned - would it be OK for you to treat these separately?" thus getting a written - and circulated - record of without offending my negotiation partner.
I most cases though it should be quite OK just to communicate openly and agree on putting the agreements in writing just to make sure that there are no misunderstandings.

Notes to self - 004, Keep local backups - The Cloud might dissolve

Keep local backups - The Cloud (aka the web) might dissolve locally

Just now a wiki that was on my list of important resources (the former wiki of the AYE conference) vanished with no clue about it's future.

It not only sadden's me because Cool URIs don't change but also because there was a lot of content even at the simple members pages, that I didn't bother to make a copy of.

The saddest thing is that it would have been easy for me to make a local copy of the wiki - after all, that is what tools like SiteSucker are for, right?

I just have to use them more often...
Therefore this very serious note to self: Keep local backups!

====8<------ End of Note to self

Comment to self:
Now the last resort for this case is the Way Back Machine, but that of course doesn't always have snapshots that include all I want to retrieve.