Because I only understood the theory of search engine technology it wasn’t as easy as I had thought to start my own and implementing the technology was an extremely complicated experience. My very first version was powered by Lucene but as it was powered by Java (and TomCat) I had great difficulty in getting it to work as expected.

I then ran a version of mnoGoSearch which was run as a Perl command line utility and while this one seemed to give me the best results I had resource issues which hampered my efforts. What I soon figured out was why Google has a few hundred thousand servers powering their search engine and after about a week of indexing South African websites I ran out of disk space. With about 8 million records indexed I had used up my 80GB hard drive.

So, in the end Grabble went PHP and while I retained some indexing capabilities from mnoGoSearch I also tapped into Yahoo to help out. At the time Yahoo didn’t have a search API so I had to do a lot of hacking to get results returned from Yahoo. In particular, Yahoo provided all the image, news, videos, blogs, sport and forum search results.

While being in awe of Google I also felt that total dominance of only a few players in the market wasn’t a good thing. I believed, esoterically at the time, that local search needed to be dominated by a local company. These beliefs I still hold true to heart and I feel that it’s critical that competitors exist. There is something very uncomfortable in knowing that one or two companies hold the majority of the world’s data and I was convinced that Grabble would be a winner.

Almost three years later and I’ve been debating what to do with Grabble. For the last two months the site hasn’t even been operational (well web search anyway) and while I didn’t want to close it down completely I also had to face the reality that Grabble was never going to dominate and I simply did not have the time to maintain it any longer.

Then along came 4hoursearch – a website powered by the Yahoo BOSS search API and running off Google AppEngine. I stumbled across this site by chance and realised that they had released their source code under GPL. Subsequent to this finding I have now moved Grabble onto this Python software and my final version of Grabble has now been hatched.

What this final version does is allow me to keep the site up for those who still want to use it and at the same time it gives me an opportunity to say goodbye. Running through the Yahoo BOSS API means that I don’t have to worry about indexes any longer and living on the Google AppEngine means technology is now taken care of.

I have learnt so much about search through my efforts with Grabble and I hold it very close to my heart. While it never took off it did give me an opportunity to expand my knowledge and experience in an area I felt passionate about and the learning curve I’ve gone through has certainly helped define where I am today.

I’m going to be talking about incorporating SEO and Social Media into a company’s marketing spend and it’s a perfect topic for me as I feel very passionate about both areas. These days I find myself working with Social Media more and more but my initial passion was always search.

I’ve long felt that while Google is an awesome company and I highly respect their ethos, one company should not control so much of the world’s data. Personally I’d rather have Google own my data than say Microsoft but I still maintain vertical search is a necessity in this ever dominant market.

When I developed Grabble in 2006 I thought I was onto a winner. Sadly it never took off (it’s almost dead in fact) but what I did gain out of this project was an appreciation for search and the technologies that power it. Grabble evolved from open source Java applications, Perl and then I settled on PHP.

There’s a reason why Google have a couple hundred thousand servers collecting, storing and processing information but none of this I could understand, contemplate or appreciate until I tried doing it myself.

What’s even better than talking about some passionate topics is the fact that I’m the second presenter which takes a lot of the pressure off me to really enjoy the remainder of the conference.

If anyone is still interested in going to the conference I have a whole bunch of vouchers which give you 25% discount off the conference price. Get in touch with me and I’ll happily hand them out.

Grabble and others should be back online in the next couple hours/days…

The Debian Wiki was created solely because I needed something to help me keep track of what I was doing when setting the server up and as such it has developed into a whole series of tutorials if you will. There is a lot of documentation out there for Debian however it is difficult to make that documentation work for the new (soon-to-be) stable release of Debian 4 a.k.a. Etch and the Debian Wiki aims to fix this problem.

Every single HOWTO on the Debian Wiki has had to be tailored to work with Etch and as such I hope that others can find value when needing to setup a similar item on their Debian based server. The Server Category items on the Wiki are confusing to the layman however as time goes on I want to have a whole bunch of Desktop related items so that people who’ve never used Linux before can actually see that it’s not that difficult to operate!

So head on over to http://linux.justinhartman.com and give me your thoughts!

It would appear that my lack of revenue model on Grabble is coming up for some critism and I must admit Uno is not the first to do so. While he may be the first person to publicly state this revenue model issue I have had a few people comment on this issue on a personal one-on-one level.

The truth is Grabble does cost me a hell of a lot of money to run and operate. I have two dedicated web servers using in excess of 50GB worth of bandwidth each month and the indexing on the servers amount to well over 70GB worth of disk space. These figures grow daily now that the website is live and building up a larger database and I am already setting up a third web server to help with the load.

Of course these figures are just going to continue growing and with that comes additional cost. How will the project fund itself is a key issue which does need addressing. As I commented on Uno’s blog the concept or idea is not to make money. That is not why I started Grabble and it will not be something that I will ever be interested in. As discussed in my Profile page I have tried to make money out of the Internet before and I made a mess of that so I do things now simply for the joy of watching it grow and develop with time.

I have said these exact words to the people that have questioned my ‘business-model’ and it does seem a little odd I have to admit. At the end of the day I do have a plan to help cover the costs of this project but this plan can only start to work once I have critical mass using the service and a product that works and functions in a way that I feel comfortable with.

My plan does not include paid-for advertising on the grabble website but rather looking at dealings with corporate companies in South Africa. I will never include paid-for advertising on Grabble as long as I am involved with it.

I can almost hear you thinking that Google made a promise very similar to this but the short answer to this is that I am not Google and don’t have any investors to account to and show a ROI. I do this for me – very selfishly I might add – and I just love doing it.

Uno also brought up an issue about Google. Now there is already a misconception about Grabble in that I am not trying to be Google but rather the best local alternative to major SA search engines. I really got so frustrated with the junk that got spat out at me in search results locally that it just became easier to use Google. I would like to change that perception and make Grabble the first point of contact when looking for something local, if this is at all humanly possible.

There is so much more still to come with Grabble and I’m hopeful to have some key stuff launched by end of January. I am so excited about this project and the changes that will take place that I just can’t wait to see how it will be received.

Grabble Shopping is the first of many new add-ons to the Grabble portfolio. This new sub-product of grabble is an initiative between Grabble and PriceCheck and hopes to provide a useful comparison shopping website for all users of the Grabble search engine.

About three weeks ago I was approached by Kevin from PriceCheck to potentially do some work together and I was certainly looking to provide some time of search for shopping websites in South Africa.

I particularly liked Kevin as the two of us share a common idea about websites should provide to the end user and it was ultimately his open mindedness that swayed me to go with him on this project.

Grabble Shopping is now a powerful tool to find products online, from South African merchants, at the best available price. We have more than 15,000 products in the database to search from and while this site is still in beta testing stages it is a rock-solid platform that works.

Give it a bash at shop.grabble.co.za and let me know what you think about it.

It’s the latest addition to my stable of projects and I’m extremely proud to share it with you! Grabble (http://www.grabble.co.za) is a new South African search engine with a difference! Grabble gives you fast, accurate and AD FREE results and best of all we only search South African websites.

How Grabble Works
Grabble is a sophisticated and intelligent search engine. Every time you enter a search query Grabble compiles results, eliminates duplicates, and displays the final results in a unique manner according to a special relevance formula. This process tends to remove much of the useless clutter and meaningless results.

Our Ethos
Grabble is only focussed on one thing – you!

Unlike many of the local competitor search engines we are dedicated and focussed on delivering relevant search results without any focus on trying to make money. We believe that searching should not be overtaken by advertising as there is ultimately no relevance to what you need by being bombarded with sponsored links and paid-for advertising.

Bigger, Better, Faster!
Grabble’s database exceeds 25,700,000 pages and documents from South African websites only and we can process as many results in less than 2 seconds. It is our belief that no other local search engine can perform as effectively and still deliver relevant and accurate results but don’t just take our word for it try a search to test it! Click here for a sample test with more than 20 million results.

Our accurate and detailed search results provide thumbnail shots of each website (if they exist) as well as a very handy Quick Look window feature which enables you to open up websites whilst still remaining on the Grabble website. By using the Quick Look feature you can immediately see if a website is relevant to what you are searching for, and if not you can simply close the ‘Quick Look’ window and continue on to the next result. This makes searching faster and easier, and less frustrating!

Our Technology
Grabble utilises the sheer brilliance of some of the worlds finest search engine engineers as the backbone for a search engine that is fast, responsive and very accurate. Grabble is hosted on its own dedicated Linux webserver with unmetered bandwidth and extensive resources in our Data center which provide maximum reliability and uptime and it is because of this that Grabble delivers quality search results every time.

Let me know what you all think.