Kettle
One of my clients has been asking about statistics for his database ... being able to metric / measure more than just the sales people (and the sales people to a better degree).
He faces the same problems most small-medium sized businesses have. That is, they have a wonderful database (sometimes it's NOT so wonderful), that is designed entirely to do data input. Running reports is not necessarily convenient.
What my client is really in need of is his database (which he has), an ETL tool, and a reporting database. This would take some of the load of the reports OFF of the database, which is used 24x7, and allow users to run "canned" reports.
I've spent some time researching the ETL tools, before writing my own, and have found that
Octopus and
Kettle appear to be two of the best. We're about to start a complete rewrite of some of the reports, and Kettle will probably be our choice. It seems to be more feature-rich, and with a built-in GUI / designer, no one has to learn ALL of the API to get off the ground and going.
I'll be sure to update as we get moving forward with the project ....
The problems with Java
There are a lot. There are more problems with Java Schools .... and, actually,
Joel does a great job
explaining what they are.We've noticed over the past few years this same phenomenon. Everyone drops out of CS and heads over to MIS. And suddenly, CS gets no more government funding (oops! -- the hidden side of the game), and MIS gets a lot of funding. The CS professors scramble for their jobs, dumb down the curriculum, get students back into the program, passing classes, and the funding comes back.
Solution? Keep the students flying through the easy classes, leave the "hard" classes for graduates (but that usually doesn't happen either), and everyone keeps their jobs. All the students graduate (even the ones who shouldn't have a degree), and we all feel good about ourselves.
Oh ... wait. Isn't that what we did with public schools too?
Resolutions First Application
We're probably about 3-4 weeks away now from our initial project release, and we're working on some of the "details". Like a name. A hosting service. Pricing. We'll work backwards.
Pricing.Most of our
competitors are non-web based, following the "old" software model. They charge somewhere between $299 and $15,000 for a software product, and the customer just "hopes" that it will work out for them. Our few web competitors charge between $20 and $50 per month for access. Which means that they get online access, and over a year, they could have purchased two or three of the stand-alone products. This doesn't make sense.
My knee-jerk reaction is to offer a few pricing levels. Specifically:
- Free. As in Beer. We're having a toss-up as to whether to limit this to certain functionality, or to make it wide open. We're also having a hard time deciding whether or not to put ads on the free site. Probably NOT Google ads (since it would only show competitors).
- Low monthly ala-carte. Pay for what you want. You want feature XYZ ... great! Base price of some low $$$ figure (say, $2), and then add on the features you would like.
- Fixed price with support. Users, some of them, really like the old-software model. Pay $199 for QuickBooks, and other than upgrades, you are set for life. Why can't we offer web software this way? Some moderate $$$ figure (say, $299), and then an annual support and upgrade cost (say, 15%).
Although I like the above strategies long-term, our first release is going to be minimal functionality. Just the basics. And we like it that way. So, how do we move someone from Excel to our service? We're not sure yet, but we do like the idea of FREE all the way. For now.
Hosting Service.Over the years, I've worked with companies (and had companies) with in-house hosting, outsourced, offshored, clustered, and ... well, every possibility. And, while they all have their merits, there's not reason to pay for a T1 to be brought into our offices, and have to manage our server.
A quick call to
OLM.net (one of my preferred hosting providers) got us an eServer1, with the hardware upgrades we want, for under $100 / month. And with a mostly text-based application, I don't think we'll reach the 250GB / month bandwidth restriction for a while.
Name.The stickiest part of our current set of problems is a catchy name. A name that makes sense to our brick and mortar customer. And yet, one that we don't lose in the sea of competitors.
Let's just say that our application is a .... Donut point of sale system. It's not, but, for pretend, let's make one. All of our competitors are "Donut POS", or "POS Donut", or "Donut Pro", or "Fast Donut POS", or "Optimal Donut", or "The Donut POS", or "POS Donut Pro", or .... you get the idea.
We'd like to stand out .... and that's proving
difficult.
Venture Voice
Podcasting has been an "up and coming" kind of media that's attaching to Web 2.0.
I've been reading a few, most of which are about ... podcasting. I think that's the same as making a program, about programming.
However, I am finding that most of the information over at
Venture Voice are great, and probably worth listening to. My current favorite are the interviews with
Joel Spolsky and
Jason Fried.
Been a Long Time
It has been a number of months (almost three) since I last posted anything. I apologize to my diligent readers, but life has been more than crazy. And, anything I needed to say was already being said by every other Blogger out there.
A few things to note:
1. We have moved to a new location, at
5102 29th Drive, in Lubbock, TX. So far, it's only been an "OK" move. I pondered purchasing the building from the landlord; it's valued at about $266,000. However, he claims it's really worth more than $425,000. My friends in the real estate business say that this is hog wash.
2. Just prior to our move, we hired another intern ... a high school senior who has the makings of an exceptional developer. Though I can only get him about 25 hours, per week ... this is sufficient for my needs at the moment.
3. We have moved to utilize
Spring in some new applications we are working on. So far, the developers are happy. I personally am just now getting up to speed, and they've already moved onto looking at possibilities with AJAX. We're looking at possibilities with various AJAX libraries. And, the
Thinkcap JX product is looking very interesting.
4. We quickly looked at, and then discarded, Ruby. Though it looks very good, we don't think it yet has a critical mass of various add-on products that it might need (like business rule processing, etc...). We'll stick with a Tomcat / Spring / Hibernate stack for now.
5. Two of our developers are neck deep in a new online scheduling application, which is looking to be produced sooner, rather than later. And we're VERY excited about it.
6. We're very seriously looking at outsourcing development of a few of our other projects, and focus more of our efforts on the requirements / design / sales of the product.
ELance.com comes to mind.
RentACoder.com seems to be smaller projects. Anyone else have any good ideas?
I've been spending most of my "consulting time" with our main client in New York, training him, his employees, and his sales consultants on how to
work smarter, not harder. Also, a fair amount of time has been spent on analyzing his (what appears to be) unique workflow. At a second glance, it's quite generic, and we're hoping to be able to apply the same processes to a more generic product available to SMBs (not Samba!).
