Marchness
I haven't written since Jan 1. I been busy. OK!
The contract at Well Fargo expired on the 27th of January, so I have been a gentleman of leisure for the past month. This translates out as I spend a couple of hours in the morning working over the job-boards. I am finding an average of 1.3 positions each day that I can truthfully apply for, and maybe one that I can stretch the resume a bit to fit. I have done a lot of different things in the last thirty-mumble years, so I have a lot of experience that kinda fits what they are looking for.... Afternoons are taken up sorting books and cleaning up my portion of the communal office, installing new software (everyone is now up to WinXP on the personal machines, and two of the Linux boxen are on SuSE 9.2.), and catching up on my magazine backlog (I am only three months behind on the Scientific American, current on Login;, current on the BMW Owners News and On The Level. I even managed to walk the dog down to the post office at least once a week. This regimen lasted for about a week.
bearmum is also out of work. She went to the NorCal Computer measurement Group's meeting the first of February and came back with a interesting proposal. One of the gentlemen she has worked with in the past was putting together a "flying squad" of mainframe performance analysts to handle a contract; a review of the compter health of a data center off shore. Would She be interested? Does the Dog Have Floppy Ears?!
The project is all dial in and work remotely, for variously reasons. The primary one is security, the Customer does not allow anyone other than their own employees on the machine. All of the work has to be done off site. There is a small mainframe in Phoenix that bearmum is doing the analysis on. The data is down loaded from the Customer Site to USB drive and then FTP'd from there to the Phoenix machine. But there were problems.
The primary tool for doing mainframe performance analysis is called MXG -- Merrill's eXpanded Guide to MVS Performance. MXG is written using a statistical analysis package called SAS. A SAS license is $30,000 per year, they are not interested in supporting small-time users, obviously. So bearmum was going to have to find something else to use to extract and massage the data so that pretty charts and graphs will happen. bearmum points out to the Boys that 'I am not a programmer'. I point out to her that I am, just a little rusty on the mainframe tools. (It has been over ten years since I fired up a TSO login and dropped into ISPF edit. But it's like bicycles, the fingers still remember which keys to push.) On Feb 4, I came on board as her assistant, the Code Jockey.
Starting on February 5th, I have been running flat-out non-stop -- trying to get up to speed on an old IBM scripting language called REXX, remembering all of the old ISPF commands, and trying to figure out ways to dissect some truly Baroque data formats. I am having a ball. My head hurts. And I keep running into things in REXX that are so simple in Perl or Shell.... On the other hand the REXX stem-variable is right cute and powerful. Did I say I am enjoying myself?
In three weeks I managed to come up with the data reduction to get the file usage spread sheet (and associated graphs) produced. Now I am working on the data from the CICS transaction monitor. This is likely to be a bit more of a Bear, since the data format is both variable and complicated. Each data-element starts with a pointer to a Data Dictionary entry that specifies how the rest of the fields in the data-element are to be interpreted. I have to figure out how to parse the Dictionary before I can begin to parse the data. Add to that, there were about 200,000 CICS records cut on the day that we are examining, and I only care about 2 or 3% of them. (First rummage through the haystack looking for the needles. Then read parts of the needles and extract the values of up to 24 different variables. Summarize and Report back before Tuesday, last, if you please.)
I am getting back into the Groove. Get up at 0800, poke around the job-boards, have breakfast and a shower, about 1000 log into the Phoenix machine and start the hacking run. Somewhere around 2200, take the dog for a walk, have dinner, finish up my status report and go for another walk while my mind slows down. Bed around 0000. Repeat.
My big data extractor just finished, and I need to get back to the mines. And I should find my copy of Oak Ash and Thorn's White Collar Holler
The contract at Well Fargo expired on the 27th of January, so I have been a gentleman of leisure for the past month. This translates out as I spend a couple of hours in the morning working over the job-boards. I am finding an average of 1.3 positions each day that I can truthfully apply for, and maybe one that I can stretch the resume a bit to fit. I have done a lot of different things in the last thirty-mumble years, so I have a lot of experience that kinda fits what they are looking for.... Afternoons are taken up sorting books and cleaning up my portion of the communal office, installing new software (everyone is now up to WinXP on the personal machines, and two of the Linux boxen are on SuSE 9.2.), and catching up on my magazine backlog (I am only three months behind on the Scientific American, current on Login;, current on the BMW Owners News and On The Level. I even managed to walk the dog down to the post office at least once a week. This regimen lasted for about a week.
bearmum is also out of work. She went to the NorCal Computer measurement Group's meeting the first of February and came back with a interesting proposal. One of the gentlemen she has worked with in the past was putting together a "flying squad" of mainframe performance analysts to handle a contract; a review of the compter health of a data center off shore. Would She be interested? Does the Dog Have Floppy Ears?!
The project is all dial in and work remotely, for variously reasons. The primary one is security, the Customer does not allow anyone other than their own employees on the machine. All of the work has to be done off site. There is a small mainframe in Phoenix that bearmum is doing the analysis on. The data is down loaded from the Customer Site to USB drive and then FTP'd from there to the Phoenix machine. But there were problems.
The primary tool for doing mainframe performance analysis is called MXG -- Merrill's eXpanded Guide to MVS Performance. MXG is written using a statistical analysis package called SAS. A SAS license is $30,000 per year, they are not interested in supporting small-time users, obviously. So bearmum was going to have to find something else to use to extract and massage the data so that pretty charts and graphs will happen. bearmum points out to the Boys that 'I am not a programmer'. I point out to her that I am, just a little rusty on the mainframe tools. (It has been over ten years since I fired up a TSO login and dropped into ISPF edit. But it's like bicycles, the fingers still remember which keys to push.) On Feb 4, I came on board as her assistant, the Code Jockey.
Starting on February 5th, I have been running flat-out non-stop -- trying to get up to speed on an old IBM scripting language called REXX, remembering all of the old ISPF commands, and trying to figure out ways to dissect some truly Baroque data formats. I am having a ball. My head hurts. And I keep running into things in REXX that are so simple in Perl or Shell.... On the other hand the REXX stem-variable is right cute and powerful. Did I say I am enjoying myself?
In three weeks I managed to come up with the data reduction to get the file usage spread sheet (and associated graphs) produced. Now I am working on the data from the CICS transaction monitor. This is likely to be a bit more of a Bear, since the data format is both variable and complicated. Each data-element starts with a pointer to a Data Dictionary entry that specifies how the rest of the fields in the data-element are to be interpreted. I have to figure out how to parse the Dictionary before I can begin to parse the data. Add to that, there were about 200,000 CICS records cut on the day that we are examining, and I only care about 2 or 3% of them. (First rummage through the haystack looking for the needles. Then read parts of the needles and extract the values of up to 24 different variables. Summarize and Report back before Tuesday, last, if you please.)
I am getting back into the Groove. Get up at 0800, poke around the job-boards, have breakfast and a shower, about 1000 log into the Phoenix machine and start the hacking run. Somewhere around 2200, take the dog for a walk, have dinner, finish up my status report and go for another walk while my mind slows down. Bed around 0000. Repeat.
My big data extractor just finished, and I need to get back to the mines. And I should find my copy of Oak Ash and Thorn's White Collar Holler