www.nathanbower.com
The Landmap Project


You can see the website here.

After I graduated from university I was approcached by someone who had seen my third year project in operation to develop a website for the landmap project.I had to create from scratch, a means of allowing users to authenticate with us and downlaod the satellite data that was the project's focus. Transfer and User logging was required, and we needed to be able to control acces on a per institution and per user basis.

I joined the project five weeks before the website would be demonstrated at a Remote Sensing conference, and work had yet to begin on the website. To make matters worse, the projcet leader would not be present for three weeks immidiatly prior to the launch! I used apache and perl as the webserver and scripting language. I converted over a thousand randomly formatted pages to a common format, designed and implemented methods for selecting which tiles / scenes / image layers of data would be downloaded, interfaced the system with the UK's academic authentication providers (athens) and coded the data logging systems. This was aside from my other duties of creating A0 posters when a project partener decided they could not do them TWO DAYS before the launch. You can view one of the posters I made (there were 4 IIRC) below.

poster

Soon I was doing everyones posters for them! I also created a set of tools to help people using the Sun E6500 supercomputer that we had, mostly in perl. I wrote a CMS for the main MIMAS website (which was slapped together IMO, and had massive inconsistencies throughout) but I left before they had chance to approve its use...br>
Here are some of the utilities I wrote:

autotar.pl This prints commands to tar and gzip the list of files that you pipe into it. The server had 23 processors, therefore you could perform 23 gzip in parallel if you send processes to the background using "&". This application simply saves on typing
html_cleaner.pl When one of the MIMAS websites was attacked by nimda, I wrote this to remove the strings causing the problem from the html files - before a patch became available!
name_converter.pl This changes lat-lon style filenames into OS grid style filenames. It is usefull when you would have to do this by hand on a regular basis!
scanner.pl This recursivly scans for a regular expression in all files below the given point of the file structure. Usefull for finding spellnig mistakes etc.
TSP.pl Similar to the above script, this perl recursivly substitutes one regular expression with another. This makes massed updates, say to a site wide menu system much esier, as the alternative would be to edit 400 html files!