Archive for February 2008

So you want to start a successful open source business? (notes, from the Paul Fenwick talk)

So, today was the OSIA Melbourne meeting. It was most fun. Met lots of people, had interesting conversations. Kudos to Donna for organising it, and all those who came to dinner and drinks, much fun was had. I wrote notes down from the talk, but it looks like Paul Fenwick and Jacinta Richardson had already written a paper on Starting an Open Source Business. They both run Perl Training Australia. Not one to throw away notes, the following are my random scraps from the talk. Also to note, Paul is a most interesting speaker!

Paul’s elevator pitch? All his course notes (and talks) are available online. I did this for my OpenOffice.org stuff, so this kind of thing, greatly interests me. Kudos to Perl Training Australia, for having such foresight.

Let the notes begin…

Paul isn’t covering moonlighting, big business (where you have VC money). The focus is an open source consultancy (small business, easy to start up).

Systems administration is covered. You don’t even need to write open source software. Training? Support?

Selling FOSS — most customers don’t care if its FOSS or proprietary — they just want to get the job done. So its easy to pitch to small/medium scale business. The low-up front costs is useful, and think about the insurance aspect (i.e. Others can work on it, even if you exit the industry).

His first business, at university, was there to support his hobbies. Just odd jobs for IT buddies, some Html, and C under Linux. Then he learned Perl for Web/CGI, and before he knew it, he started his own business. As a sole trader, he was also a university tutor, and enjoyed tutoring, and the pay was peanuts in general. Eventually he got a real job, and he realised that having a real job sucks! (long hours, commute times, boring work, deadlines, not learning new things)

He then looked at consulting friends, and they earned $100/hour. He then thought, to earn $50,000 a year, he’d only need to work 2 hours per day (!). So, he started a second business…

Lesson 1: Money. Start personal accounting, and use GnuCash. If you want to start a business, start doing this now, because you want to know your financial status instantly. It makes taxes easier. You can also profile expenses — know what you’re spending on is crucial (know your savings longevity).

When you start your business, your savings will dwindle. Dwindling savings is scary, and you might be scared back into a real job, resulting in you wasting time and money.

Business with friends, was great. Meetings by pool, over BBQs, making beer, etc. Business focused on consulting, but he wanted to teach.

Lesson 2: Business with friends. When there is irreconcilable differences of opinion, you tend to lose friendships. Imagine, losing time, money, and friends!

He then started Perl Training Australia, as business number three.

Lesson 3: Contacts. Contacts from university, work, contacts (friends of friends). Most work came from word of mouth, as opposed to advertising. This is good, its cheap advertising. How do you get contacts? You network. There was one stage when he was cornering people on the bus even — he just talked to everyone. 30 seconds of what you do, what your skills are, why you’re important, and give them a business card.

So, you need money and contacts. What about talent? This is optional ;)

Lesson 4: Social Skills. What you really need is social skills with non-geeks. Your clients, don’t know if you write good code. They just want solutions! The only way your client knows you’re doing a good job, is if you’re telling them. Give them a daily report, if need be. Effective communication is important.

Most people don’t use the golden rule of social skills. Don’t be unwarrantably honest, don’t be insulting, be creative. Appreciate others, be polite. Don’t cause anyone actual hardship.

Discovery: clients were also small to medium sized businesses. They appreciate ‘personalised service’. No one likes going through the helpdesk loop. Have penalty rates before hand — imagine getting called at 2am on a Sunday morning!

Remember, there are different types of clients, and the difficult clients produce the most work. Ideal clients? Technically clueful and well-resourced. They hire you for specialist skills. Unfortunately, you won’t get too much work — too clueful and organised. Call them regularly, make them feel loved, so remain fresh in their memories.

Difficult clients, tend to be disorganised, have few technical skills, and may need full-time staff. Everything is always an emergency with them!

Intermediary clients — these are people that want you to do work for their clients. These folk can be interesting to work with — they can do specifications, scoping, management & client liaison, and even invoice on your behalf. They can also be awful — imagine getting poor specifications and scope. Not having access to the real client, you have to wait ages to get an answer. Worst? “You get paid, when we get paid”.

Most satisfying work, tend to have good closure. Some things never have good closure — this is like system administration. Its never going to end! Software maintenance, doesn’t have good closure (features, documentation, etc.).

Poor closure = good income stream!

“As a consultant, I can choose the work I do.” However, what it really means is that you choose between work and starvation! You cannot be picky at the beginning. You will have to do grunt work, before you gain the reputation.

Sys-admin and development work during the start, was hard for Perl Training Australia. He couldn’t develop training materials, and develop the actual training. He left his full-time work to gain more time, but during the first year of the business, he was working 12 hours per day! 2 hours per day, is clearly not enough. Clients want their work done now, not by Christmas.

So he worked 60 hours/week, and was happy. He kept saying he would take time off later. Remember, clients need constant support, so you’re not going to go on holidays.

Consultants get heaps of money. But only for billable hours. They don’t include things like office tasks (getting stamps!), book-keeping, travel time, systems administration of your own systems, meetings with clients that don’t employ you.

Ideally, you would like to make money without spending time. This is probably the number one reason to start a business. Good way of doing this? Recurring revenue via a hosted service or a subscription service. In training, each extra person on a course, don’t represent a big cost. Product sales is also useful.

Employees do work for you. They spend their time, to make you money. This means you can then do more work, or grow the business.

Hiring employees is a lot more work. There are legal aspects (tax department, contract negotiation, etc.). There is the requirement of office space. You need to show them the ropes of the business, because they will not know how your business works.

When you’re busy, it is the wrong time to hire. Think of the Mythical Man Month. You’re going to be overloaded. The right time to hire is when things are quiet. Always, have foresight, and planning as a small business owner!

In a large business, there are differences between sales and systems administrators. In a small business, the positions become less well-defined, so you’re going to have to ensure you and your staff multi-task. University students don’t really have this “Small Business Shock”.

Keeping the perfect employee? Keep them happy. Spoil them. Do whatever it takes. Because remember, hiring someone new, is actually expensive; you’ve already trained your employee!

Technorati Tags: , , , , ,

How do you check the speed of your Ethernet card?

It has been a very long time since I looked at the LPI (training, courseware, materials, exams), and I don’t remember if this is even covered in LPI 101/102 (and I have no idea about Level 2 or 3, I’m only Level 1 compliant, and even then, back in the Red Hat 7.2 study days :P). Either way, I hope this note is useful. So, how do you check (or change) the speed of your Ethernet card?

Quite simply, with the help of ethtool. Some might be saying, why not mii-tool – well, its because mii-tool is now obsolete, though you pretty much have it on all your Linux distributions. Why? Because it comes as part of net-tools, which also provides other useful tools like netstat, route, ifconfig, arp and many more. If you don’t have ethtool installed, its available in a package, aptly titled, ethtool.

Usage is simple: ethtool eth0. Replacing eth0 with whatever your interface is, is fine. You should now be able to see the speed of your link, and if you were for instance changing it, you might do ethtool -s eth0 speed 10 duplex half or something similar (read the man page, its really helpful).

Me? I just wanted to see if I connected myself to a 100mbps switch or a 10mbps switch and was too lazy to look at the wiring closet ;)

Technorati Tags: , , , , ,

When OpenDNS is unreachable

Beware, whom you trust your DNS to. I have particularly enjoyed using OpenDNS, mainly because I travel a lot, and encounter a lot of broken DNS configurations. However, when OpenDNS goes wonky, it makes you really wonder if running local DNS makes more sense.

[-(byte@hermione)-(pts/1)-(11:57pm:27/02/2008)-]
[-(~)> ping www.google.com
PING google.navigation.opendns.com (208.67.219.231) 56(84) bytes of data.

--- google.navigation.opendns.com ping statistics ---
7 packets transmitted, 0 received, 100% packet loss, time 6000ms

Clearly, not good? Quick edit of /etc/resolv.conf, and I’m quite happy to use local DNS. Oops, it seems like OpenDNS is clearly down now (can’t even reach opendns.com). Ironically, only from my Unwired Internet account.

The difference in traceroute output (for status.opendns.com):
12 38.104.140.46 (38.104.140.46) 261.618 ms 267.483 ms 268.207 ms
13 208.67.219.60 (208.67.219.60) 220.762 ms 220.735 ms 220.709 ms

From hop 12, to hop 13, it is unreachable. Let’s hope this temporary problem, disappears.

Update: Its back online, 1.45am.

Technorati Tags: , , ,

Some thoughts and transcripts from the Alan Cox video series

While catching up on some interesting chat on IRC today, I decided to watch the Alan Cox video series that Red Hat Magazine recently placed online.

In Alan Cox and the state of free software:

  • Alan speaks about software patents, as a problem for free software. Lots are starting to understand that they don’t work and they violate international treaties.
  • Alan talks about political systems – so you can’t get free software into government or schools, because of certain vendors that they choose. Approved suppliers cause grief, when they only supply proprietary software.
  • Alan talks about the OpenDocument Format and OOXML mess, and it confuses people, who want standards (FUD).
  • A challenge now, seems to be that there are a large number of free software users now, who are not technically adept. So software has to adapt to their requirements – people don’t apply patches, etc. anymore so you just have to find new ways and tools to deal with bugs, etc. We in the free software community have to scale.
  • Liability, with regards to poor quality products. Security software for instance, has this problem. The free software community has to work with liability law, as governments will increasingly require it. What standards/expectations should free software hold up to?

In Alan Cox on community and the enterprise:

  • Alan speaks about his history at Red Hat, and his history with Linux and how he began hacking on it.
  • A subscription model, and Mark Webbink figured this out – he built the model, while staying true to free software. If Red Hat hadn’t done that, they’d have lost a lot of community support. Alan focuses on how having a community distribution in parallel made sense. I can’t imagine if he means Fedora and RHEL, or CentOS and RHEL? :)
  • He was referring to Fedora! Red Hat is just helping, enabling, providing resources, guidance & advice – the right way to work with the community. Everything else is generally community controlled.
  • In developing markets, open source is important – buying local resources to create/maintain it – keep the sovereignty.
  • “At the right place at the right time”

In Alan Cox on the kernel, patent promise, and the progress of free software:

  • Late 90’s, everyone jumped on the open source bandwagon without any idea what it was about. This was the same period where you stuck an “e” in-front of any company name, and it became a new business model. How true.
  • The 2000 crash, helped the serious companies work open source better. Oracle for instance didn’t take Linux seriously – now its an important part of their business model.
  • Alan talks a lot about how he now spends time cleaning up code.
  • Red Hat, at the end of the day, have a patent portfolio. Nobody particularly wants to launch any, but if you don’t have any to launch you have a problem – kind of like nuclear warfare.
  • Software patent promise: people doing free software will never have to worry about software patents. People doing proprietary software, it gives Red Hat a mechanism in case of patent lawsuits (also, to trade patents). No one knows if how well this works, yet. Just hope this nightmare, never happens.
  • Who knows, that in the US, the Supreme Court might decide that software patents are not valid? No one knows, since no one has pushed it this far.

Technorati Tags: , , , , , , , ,

Registered to be a postal voter…

Today, me, and about eighty other folk, went to Malaysia Hall, to register ourselves as postal voters.

From what I understand, the forms we filled out need to reach Malaysia by 5pm (UTC+8) on the 27 February 2008. Assuming that the official close of 5pm (UTC+10, +1 DST) is met, this leaves 3 hours to fax all the forms to Canberra and to the Election Commission in Malaysia. This is going to keep one Dr. Ali (Education Attache, Malaysia Hall, Melbourne – since we don’t have a consulate in Melbourne), rather busy. He’s also going to courier the original forms back.

One can now only hope that in the next week, a postal voting form appears. Everything was professionally done, only the identity card (IC) was required (no passport, as initially thought), there were sample forms that were filled out, and there were quite a number of people there when I was filling out my form.

The talk? Keadilan (PKR) [they also deserve special mention: minimum wages in Malaysia], PAS, DAP, and an independent candidate! Don’t know if our minuscule Melbourne votes are going to make any difference in this election, but here’s hoping.

Again, mad props to Tirath for helping sort this out. Also met up with an old friend, Raja Azlina while I was there (from MIMOS days, its been about a decade or more since we’ve known each other’s work).

Technorati Tags: , , ,

Absentee (postal) vote registration in Melbourne for Malaysians

Tomorrow, is a big day. If you’re a Malaysian in Melbourne, and are eligible to vote, the absentee (postal) vote registration happens tomorrow (27 February, 2008). Short notice? Apparently, the embassy in Canberra didn’t even know the deadline was to be tomorrow, until a phone call from a certain someone, Tirath, was made today (26 February 2008) at 4:45pm (UTC+10, +1 DST).

Event blurb:

Date:

Wednesday, February 27, 2008

Time:

8:00am – 6:00pm

Location:

Malaysia Hall Melbourne

Street:

K4, High Street, Windsor VIC 3181

City/Town:

Melbourne, Australia

Tomorrow is the ONLY CHANCE to register. Call Malaysia Hall before you go, just to double check that they are open and ready (04 3343 7263 / 03 9529 4507). If you come, make sure you bring your IC and your passport.

Windsor is easy to get to. There’s public transport (a tram stops at Malaysia Hall, or nearby, afaik), and there’s a train station as well. Mad props to Tirath, I hope lots of people go, and register, and vote (the opposition, nonetheless ;) ). Facebook event, if you can see it.

Technorati Tags: , , ,


i