2014-10-04

Is Software Architecture a Service, Needed by Small Companies in Kuala Lumpur?

So Ispecificaseen the same problem in four software projects I've worked on over five years. Lots of developers everywhere. No architects. Every week someone comes up to me and asks for software development services like headless chickens. I'm getting over this - f.k.oherent with flexible / fail-fast / iterative / agile work-flows and culture: whatever conventions you're relying on, you can derive more efficiency from a clear set of blueprints and a whip. I just pitched back to on RFQ that I'd be up for software architecture services, on a long-term low retainer:

1. Focus primarily on usability concerns,
2. manage the critical chain,
3. pick the right implementation libraries, tools and methods & manage choices on an on-going basis, and
4. site-manage the development/dev-ops (* arguably the most important bit).

Pretty much what architects do in brick-n-mortar projects.

I don't know if there's much demand for this in the market, but I think the positioning of my service layer is clearing up a bit now.

Arguments for/against this approach would be appreciated.
It's quite interesting. As a profession, software architecture is a bit stuck in the 70s with flow control diagrams.

Civil architecture is a 4-8 year liberal arts and sciences degree requiring calculus and physics. (In some places.)

I wonder how long it'll be before software architecture becomes a thing, like civil architecture is a thing. ("thing" reads: "social norm")

Later in the month:
Can't hire a CTO? Maybe a full-time employee is not in budget.

If this is your situation, would you consider partnering a software services company to fulfill the CTO role in your company? (Scope: usability + commercial concerns -> technology cost/scalability strategy -> development architecture -> project management of development.)

Possible duration of contracts: 2-5 years to ensure continuity.

Possible price of retainer: as low as RM2,000/month, depending on requirements.

Where I'm going with this - I'd like to start a firm. I tend to think of this set of services as being close to "architecture," the way civil architects work on realestate projects. It's an experimental model. I'm looking for traction. To discuss the model, please post in this thread: https://www.facebook.com/jerngatwork/posts/336456446535069

To discuss actual engagements, please post below. (I'm preparing for a lot of F***Us, lol.)
Let it be noted, that there are companies on the market that have been working in a consulting capacity and matching CTOs to startups. I don't think that's what I'm curious about, at the moment. Also it is not good for an architect to have an overly broad scope of work, as the business of knowing how people interact with technology, and how to make a net profit from this, is already rather broad.

#meatShop #scopeSalad

In our industry, the business users usually know what they want, and often hire a series of staff to implement it (often by accident, since turnover is so high). I'm just sick of the situation where in three years a company has to build *the same features* three times because different people disagree on the implementation, and the company has no long-term view. Saw plenty of money burn this way a few years ago. Still seeing money burn. Not happy.

//

And then there's the business users who don't actually knoww what they want: those who have an abstract intuition, but are unable to signoff on concrete plans. Sometimes they signoff on abstract plans without realising how vague their plans are until implementation fuckups occur. So then who should they blame, the implementors, the architects, or themselves? It is not *always* clear.

//

I'd prefer it to be scoped within "architects" for hire; but someone told me CTO-absences would be a good copy/channel strategy. I don't want to displace CTOs, since CTOs have very different scopes in different companies. I DO want to take all the long-term responsibilities for maintaining and monitoring architecture of implementations. I DON'T want to take development, maintenance, and key operations risk (as opposed to strategic risk) responsibilities. Just testing the market here.

//

I'm quite aware that people need operational headcount, as I've done that job before here and there (not always in IT). However, I'm not particularly interested in taking on ops risk for anything other than top rates. 

I notice that your service offering is fairly broad, including some business planning and hand holding of non-technical founders. Not sure I want to go there with this project (I have delivered, and in the future would do those services under a different brand from the firm I'm exploring here). That stuff is easily defined and rather unchallenging to me at this point.

I'm specifically interested in pitching a clearer definition of the scope and definition of software architecture as a business function, for a low fee, over 2-3 year time frames at least, especially in agile teams where it tends to be neglected (debateable, of course). So yeah, internally I'm approaching an old market with a new approach. Too bad if it doesn't work out :) but NPD is the fun part of startups.

//

let me throw some number out there for context. I've worked from RM5/hr to RM319/hour in this town. I've been yelled at for missing deadlines I didn't promise, as well as thanked for REMOVING value from a project. While all of that is quite entertaining... I've generally tended to hedge my approach. (a) charge a ton to put up with anything (b) charge a little, or whatever the customer wants, and give them what they ask for even if it's crap (c) still pivoting. Heh. TLDR: my life is designed to keep running on minimum wage. Heh, most of the time though, I'll take what the market offers and GTFO everyone's way got better things to do.

2014-10-02

Two Thoughts at Bedtime

Every so often I'm reminded that I'm a technologist by career choice, and not because I care much for technology. I spend years studying the craft, because it is a discipline. Otherwise I would probably be some sort of careless creative. And then I have to listen with all the folks who say, that one cannot excel in a field where one is not impassioned. I spit on you.

Meanwhile, all around me people wonder about my country, and my continent. Indeed I live in the rectum of vainglorious democracies.



Flat tire. Waiting for mechanic. I should buy a jack.
Thoughts at bedtime not entirely accurate. I'm a career technologist in the sense that I was motivated to begin study engineering details in order to become a better manager. At the end of the day, I'm only in commerce to pay for my hobbies, and so it's really about the bottom line.


What might those be?
Information architecture. (tldr)

All the liberal arts and sciences, to know what needs to be copied. Engineering to know how to do the pasting. I'll move on to other hobbies once we have machines that are equal poets, physicists, and lovers then our meaty peers.

Client-side JavaScript Frameworks

A buddy of mine at ServiceRocket mentioned Ember, Backbone, and Angular to me a couple of months ago. I just picked up Rails, and am looking into the most minimalistic way to extend it to build apps that live mostly in the browser. So I was reading up.

The first three-way comparison that I read.

Related for comparison:

Why this guy wants to use Riot.js (1kb) instead:
- https://muut.com/blog/technology/frameworkless-javascript.html (Now, THIS is something I can agree with.)

More three-way comparisons:
- http://www.100percentjs.com/backbone-or-angular-or-ember-here-is-my-choice-and-why/
- http://blog.fusioncharts.com/2014/08/angularjs-vs-backbone-js-vs-ember-js%E2%80%95choosing-a-javascript-framework-part-2/

Broader comparison:
- http://sporto.github.io/blog/2013/04/12/comparison-angular-backbone-can-ember/

2014-09-28

I Like Julia

Finally had time to check it out, and I'm sold. I'll have to get more used to it, to know how it works vs Haskell.

Vs haskell:
- J is more team "engineering / ops" focused
- JIT default runtime
- more straightforward type system (see point 1)
- DARPA backed, so I worry less about platform maintenance if say SPJ, god bless his soul, were to disappear from GHC
- and omg, the langdocs are so much clearer (again point 1)