Scott's Column
Application Server, Telescopes & Flavors of Windows

March 1, 2002
By Scott Lewis

Never swap DLLs on a production application server. I did this and shut down production for 9 - 10 hours. Not Fun! However, I did manage to redeem myself by building a proper test server and corresponding client workstations. Also, my laptop is on its 3rd repair.. just one more to get a replacement. Finally, I started doing some research on telescopes.

Application Server

We have an application server that performs a number of functions. One of which is to be used in a distributed processing environment. This is done by running a SOAP server on the machine. The program on the client creates an XML document, or package, and sends it to the SOAP listener on the application server. The server takes the request and processes it to launch another program with parameters in the XML document. The application on the server updates tables in Oracle that are read by the client so the user interface of the client can show the progress of the application server on the given task.

It is not a completely smooth operation. And it is in a very delicate and undocumented state. The programmer before me was very clever in figuring it all out. Microsoft's SOAP server had difficulties when he implemented this so he wrote his own SOAP server (affectionately called Psuedo SOAP). But he did not finish debugging the whole process before he left the company. That's were I came in.

Part of the problem is that the client applications were modified to this architecture. Basically big chucks of their code were moved to the application server. So there are a lot of references in the code that are confusing as to where processing is taking place.

When I updated the client, the application server stopped working. I reinstalled the last "working" version of the software, but the client still didn't work. I traced the problem to a DLL that was not running properly on the application server. I assumed that the DLL being called by the application needed to be updated because of other dependency changes to the client. So I re-compiled the DLL and copied it to the application server. No change. So I copied the old DLL back and went back to the client. Eventually, I installed the most recent production package for the client and installed it to another workstation. That still did not work (but should have... as you'll see). Eventually my boss re-registered the old DLL on the application server and everything started working

This means that only the client was at issue. Installing the production package to the second workstation should have fixed things. So that was my screw up. I still had dependency issues as to why the first client workstation still could not run the client software.

New Application Server

All this could have been avoided if I had a proper test/development environment. So we "stole" a computer out of the classroom and I loaded it with Windows 2000 Server to re-build the application server in a controlled environment. I took a very tedious approach to building the new server... I call him Junior. Anytime I had to tweak something in an attempt to get Jr. running, I would reformat his hard drive, and load everything from scratch... making sure to incorporate any tweaks into the load procedure.

It took about 4 or 5 reformats to get it all working perfectly. But it was well worth the time and effort. I can now rebuild the application server in 2 or 3 hours. Also, Jr. can act as a backup to the main application server giving us redundancy... something that would have done a lot of good the day I brought production in to a screaming halt.

While I was rebuilding the application server I also rebuilt the client workstations with the same vigor. I was able to eventually determine why that first client stopped working properly. I now have reproducible and fully documented procedures for installation packages that can be used at anytime to build a replacement workstation or server at a moments notice.

My director and supervisor had me burn all the packages and documentation to CDs so we don't have to worry about them getting lost on the LAN somewhere. I expect to make a few copies, one of which will go offsite where we keep our backup tapes.

Telescope Research

As mentioned last month I may try to sell my Desert Eagle 44 Magnum to fund a telescope purchase. I received grief from my best friend (in NY, or is that NJ) that I didn't ask him about telescopes before putting an open question in my column. He has been an astronomy buff for a very long time. I guess I didn't ask him because he almost never talks about it. I also assumed that he may have been out of touch with current equipment having had a good telescope of his own for a very, very long time. Sorry buddy, I should have asked you first. 

My friend gave me a great recommendation. But I wanted to know more than just a great model to buy. I also wanted to know what it all meant. I couldn't make heads or tails of any of the information in ads, so I started doing some research. Unfortunately the more I learned the less I felt I knew.

Types Of Telescopes

There are three basic types of telescopes. Refractors, Reflectors and Catadioptric. Refractors are straight through telescopes that use optics (lenses) to bring objects into view. Reflectors use a mirror (or mirrors) to collect light and reflect it to an eye piece for you to view. Catadioptric use some combination of optics and mirrors. Within these categories lie some complications. I will try to present what little I have learned about each in a quick paragraph.

Refractors: Most refractors use two or three lenses in a typical long tube. Less expensive (not cheap department store ones) refractors use two lenses and are called achromatic, others use three lenses and are called apochromatic. Usually these telescopes are the most expensive for there size (size being the diameter of the lenses, also called their aperture). It is the aperture that is the most important part of a telescope. Assuming all else being equal (quality of optics and mirrors, etc.) the larger the aperture the better its ability to bring light into view. Lenses are more expensive then mirrors to produce, so refractors get very expensive when the aperture gets large. (I will not concern myself with the cheap telescopes sold at K-Mart here.) Apochromatic telescopes use the third lens for color correction and to help with other anomalies.

Reflectors: There are a number of reflector telescopes, but the most popular is the Newtonian Reflector. A Newtonian reflector has the eye piece at the front (top) of the telescope. Light enters the front, reflects off a large mirror in the bottom of the tube to an angled mirror in the center of the tube at the front. This small angled mirror points the image out the side of the telescope to the eye piece. The Newtonian gives the most aperture for the money. Like I said, mirrors are cheaper than lenses. Dobsonian is a twist on the Newtonian reflector that puts the tube into a "bucket" rather than on a bulky tripod. This makes it easy to adjust even with large tubes, and these bucket mounts are fairly inexpensive.

Catadioptrics: These use a combination of mirrors and lenses as part of their optics. These are called Cassegrains. They can be broken down into two main "sub types," Shmidt-Cassegrains and Maksutov-Cassegrain. Both uses a lens in the front of the telescope that focuses light to a rear mirror, light bounces of this large mirror to a second, smaller mirror in the front of the telescope, and finally the light is reflected to the rear of the telescope where the eye pieces is. The main difference between Shmidt and Maksutov is the type of the primary lens at the front of the telescope. The Shmidt uses a lighter lens making the telescope lighter and more portable than most other telescopes. This has given the Shmidt-Cassegrain a huge following in amateur astronomy.

What to buy. My friend suggested the Celestron 114GT, and it is a great choice. He found it for $329. A bargain for 114 mm (about 4.5") of aperture. It is a Newtonian Reflector scope with a "Go To" feature that allows the telescope to point itself at something like 1,400 objects in the sky. I have not been able to find a scope that would be better than this for under $500... assuming the added cost of a GoTo feature.

However, in my reading I have found that I may want to go slightly over my original $500 budget. I have read reviews that seem to put a lot of excellent scopes in the $500-$600 range. And from what I can tell it will end up being a scope I will grow into very well. Unless, of course, I end up giving up on this hobby and the scope ends up collecting dust in the garage instead of light from the sky. Then spending a little over $300 is better than spending almost $600.

I am having a hard time deciding on what I want. It is hard trying to determining exactly what comes with a telescope for the price. Some of the $500 scopes don't have tripods, others only have one eye piece. Newtonians are open at the front so they require maintenance to keep them clean and their mirror aligned, etc. Shmidt-Cassegrains are closed and mostly maintenance free, but cost more for the same aperture. Dobsonians offer a huge leap in aperture for the price, but don't have motorized mounts and don't have "GoTo" features. What to do?

I am still researching. Also, I have yet to try and find anyone local that can "introduce" me to astronomy. Stay tuned... I will have more.

Windows XP vs. Windows 98

If you read my Feature Article this month you will get all the details of my "living" with Windows XP for about 5 - 6 weeks. I gave up when trying to get Return to Castle Wolfenstein working caused my system to get hosed. I never got Wolfenstein running under XP. It ran first time out under Windows 98, and the machine is considerably snappier in performance to boot.

I would like to restate here that I ran Windows XP with only one crash in the 5 - 6 weeks I spent running it. I have to reboot Windows 98 every two or three days. I would like to try Windows 2000 Professional at home, but I don't have the time at the moment.

Server At Home

Ideally I want to build a true server. It would be simple, mostly a file and print server. But it would be a dedicated machine. That would leave me the freedom to do what I want with my personal machine. I have a few spare parts laying around, and may try to work with them to build something. Money is tight, and I will have a hard time justifying the cost of two machines. My current machine is slated for the kids, so I won't be able to use it for a server.

I have been do some creative thinking on this server/workstation and will report on that next month. Stay tuned...