Web Lab and FTP Links

Webber is an Apache Server at Sooty Solutions. Since Webber is under Sooty's control it can support more sophisticated applications.


Webber's Installation and Configuration Notes

If you want to know how to set up a web server in your basement the guide is here.


Webber's Information and Services Test Page

Visit Webber's experimental services here.


Webber's FTP Services

If you want to exchange files (incoming or outgoing) you can use one of the following FTP guest accounts. Of course you will need to get a password from James or Dean in order to access these FTP services.

Internet Explorer must be set to "passive" mode for FTP services to work. Netscape and Mozilla/Firefox work fine without having to specify the mode.

              guest1 Account.

              guest2 Account.

              guest3 Account.

              guest4 Account.


Projects:

MarketWired RSS Feed Notifier in Java

Sooty is a dividend investor and the Yahoo Finance portfolio feature used to be a great place to track investments. Over the last few years Yahoo has changed the layout and content so it is becoming more labor intensive for the serious investor. I love my broker for its "banking" functions but it does not do "news". I've tried some of the free portfolio trackers (Google, TMX, Morningstar ... you need to create an account) but the current Yahoo layout is still marginally better.

All I want is dividend notices for the stocks I own. Yahoo financial news section (below the stocks table) has become so cluttered with paid ads and news feed from "analysis" companies offering free advice or "for fee" reports that I have to scan through many pages of junk links to find what I need.

In frustration I wrote a little Java program to parse the MarketWired RSS feed for just the news I wanted. I thought it might be useful for other investors. The program is a simple example for anyone wanting to parse RSS feeds and provide and screen alerts on news items. See MarketWired RSS Parser.


Automated Backup to Amazon AWS Glacier

The Glacier Amazon Cloud Service provides a cheap way to store data in the cloud (1 cent/Gb/month). The major cost is transaction fees but a careful attention to the terms of the service can keep them low.

The java interface required to save and restore data is "bleeding edge" but with a little help anyone with good java skills should be able to get it running. The project shows how implement an automated off-site backup into the cloud at minimum cost. See AWS Backups.


Automated Backup to Google Drive

In 2012 Google started cloud services. They give away 15GB of free cloud storage for use with a variety of Google products (mail, word processor, spreadsheet, presentation slides). I only use gmail (which is a superb product) and most of that 15GB sits unused. The storage can also be used with Google Drive which allows sharing of data files in any format. So why not the space for automated backups?

The main problem is Google Drive is designed to be a manual interface. You don't need a library of services (like Amazon AWS) to access Drive but it is not as reliable as AWS. My first attempt was using a simple Java program to place backup archives in the local Google Drive folder (on your PC) and Drive would sync these to the cloud storage. That attempt is archived here as Google Drive Sync Backups.

Using Google Drive Sync not reliable: the sync falls asleep and won't wake up and simultaneous syncing of too many files will cause the sync operation to crash. The bigger problem is the error handling is hidden inside the googledrivesync.exe daemon so you cannot write Java code to improve reliability.

My 1st attempt made sure the daemon didn't stop uploading but this left two fatal failure modes. About every 10GB the upload gets into trouble and the sync process retries the upload. While the upload does succeed it consumes 3 times the archive size in network traffic. My archives are set to about 5GB (the size of a DVD) so this failure mode will cost 15GB on my internet bill and take 3 times as long to complete.

The more troublesome failure mode happens about every 20GB when the upload never finishes. I throttle my Drive traffic (on a dedicated PC) to 100Kb/second (9GB/day) but "automation" means "unattended" so a month in this failure mode will consume 270GB of my internet plan.

The 2nd attempt uses the Google Drive REST SDK. This is similar to the AWS SDK but it has a number of major issues. The libraries are built on the Apache implementation of REST (REpresentational State Transfer) and it requires many "Google modified" versions of the standard JARs.

The Google libraries replace a number of standard Java classes. The most annoying is java.io.file: during debugging your code will not compile for bizare reasons because of the conflicts. For example the first thing you learn as a novice programmer is [File fileHandle = new File("test.txt")] but this will fail because Google redefines the File services.

The Google libraries are going through a rapid evolution and there are many versions. Classes and methods appear and disappear as you try to keep up with the code churn. This means that working code will fail in unpredictable ways if you try to remain current For example early versions use "title" but later versions use "name" for the text handle of a resource (like a file name). Data structures constantly change so your "perfect" design may no longer work because the data structure no longer exists.

This 2nd attempt needs an experienced Java programmer (a bottle or two of Makers 46 helps). But the package I provide does work reliably and all pieces from my example code come in a ZIP so they work together. I am using this code as the basis of an automated unattended backup system. Not discouraged? ... leap to Google Drive Backups.


Moving an Old Windows 7 Disk to a New PC

In 2014 my kid gave me a hand-me-down PC with sweet specs (water cooled quad core 3.8Ghz, 16Gb ram, Nvidia GTX 680) but my whole life was on a Windows 7 disk on the old PC.

No one moves old Windows 7 disk to a new PC - it won't boot. The answer is always reinstall Windows then every other thing you had on the old PC...but what if you could just move the disk...how much trouble could it be?

The project is not for the faint of heart. There are multiple failures and frustrations at each step but it can be done and now I have my familiar world on new hardware and it only took about 12 hours.
See Moving an Old Windows 7 Disk to a New PC.