Mail Me

Week 11, 2000 ,Svenson

Sjon



-- Previous week -- Most recent -- Next week --

Kelshon Saga. The logs. (book37.2 p83)

13-03 to 19-03

 

 
Top <<<     Mon -- Tue -- Wed -- Thu -- Fri -- Sat -- Sun     >>> Bottom
Day view


MM-lxxiii     Monday

 

2000-03-13

 

 

The forecast was for -3°C and dry. Well they only missed the sign. Nice cool and dry with a few hours of sun even. Late in the evening it became misty which kept thickening until it was indistinguishable from light drizzle.

As usual my Monday was quite productive. With two problems solved . That is from picking up the problem report over checking the error (all reports must be checked because often there are problems reported where nothing is wrong, or where the problem cannot be reproduced. These cannot be solved. ) and the actual (environmental) behaviour. Over crushing the bug , testing the program internally and environmentally (internal means checking if the adaptation works as it should, environmentally means check the rest of the program + the calling and called programs to make sure no new bug was introduced). All the way to transferring the program back to System-Test and administratively checking off the problem report.
Did you notice that the crushing was actually the shortest item in that paragraph?

I dropped the frame handling from my current and currentw pages. I never got substantiated complaints other than a few people remarking they didn't like frames (but adding that it was not enough reason to not use them if I liked them) and some people mentioning that 'some browsers have problems with frames' (but then not pointing out the actual problems). The main reason I drop them is that I notice that the main daynote frame doesn't get the focus so scrolling doesn't always work. Another is that I noticed that I don't use the upper frame much ( and the links are repeated on the daynote itself anyway. Maybe I could have checked the site statistics first but I don't know if that would help me much (and I would have to find out how to get at them first).

I think I found the perfect PC.
Dual processor? Nah.
Gigabytes of RAM? No.
Clustering? Njet.
Disks at 20.000rpm? Nopes.

Just this , of course it is called Media-Pro.

 

 

 
Top <<<     Mon -- Tue -- Wed -- Thu -- Fri -- Sat -- Sun     >>> Bottom
Day view

MM-lxxiv     Tuesday

 

2000-03-14

 

 

It is raining again, not hard and not constantly but it is definitively wet. In the mean time the temperature is up again at around 7°C.

In contrast to yesterday I didn't perform much today. Oh, I did create the NOW- and HTM_ files with a few indices so Wilbert could build the server-chema for TeleSales. Not a big job, just a little time waster.

And I patched up an OLI program for the HCC_ file adaptations. While I was doing the adaptations about a month ago Sven, from OLI, came by because he had to do something with the classification codes in OLI. I explained him how the new system with the HCC_ file was organized so he could build it in in OLI. Last week Jan noticed that the classification stuff didn't work and came to me with his remarks.
Yep It all works in OMSI-3 but not in OLI. Either Sven did nothing about it or he did make only one or two changes and left the rest untouched. And I forgot to check. He has since left the company for a new job so I will have to do the changes. The changes are not difficult, I lose most time searching which programs must be adapted than doing the adaptations.

Well Today I lost most time due to a few consecutive crashes of my PC. Most of the time it is stable with only Opera (browser), Netscape (mail) and Client Access (terminal emulation to AS/400) running, sometimes I have CuteHTML and/or Lotus Word Pro open as well. Nothing extracting, a Pentium 200 (Clasic+MMX) with 64MB should cope with it and mostly does. But today was different with lockups swallowing even the ctrl+alt+del salute.

Last week I was pressed for time and didn't read much from the other daynoters. Today I picked up on most of them (some days remain unread but that's life).

For about the first time I checked my site statistics. I don't know enough about it to understand half of the data presented by iTool and their help only tells me what I can do with the data but not what it actually means. For example what is the difference between a hit and a session or between a page load and a page read?
Well I changed from frames to redirection and I wanted to check how that would impact my site. Therefore I wanted to take a snapshot now and compare it to a snapshot in a few weeks time. And I wanted a comparison between my daynotes and my weeknotes.
I found out that nobody reads the weeknote, which stands to reason because it contains nothing more than the daynotes and it is not linked off a referring site. Week notes are easier for searching but daynotes are more 'rounded', they look better finished, a week note obviously is only finished once verey seven days. But I do have double work keeping them both.
The main thing I noticed is that the iTool statistics pages (generated by Statistics Server 5.02 from Media House) don't like being printed by Netscape and they don't like Opera either. Well I got a bunch of papers with only the URL at the top and a page number at the bottom. The strange thing is that some pages printed with data.
The other big surprise was that the 'currenttop.html' page was visited most. If you remember that is the small top frame containing nothing but a few links. The next most visited page was the current.html page that defines and presents the frame with the daynote. Its frequency was about half that of the meaningless top frame. Now that I drop the frames I may get some better stats in stead of overblown pages. Although ...
Another strange thing (error?) is that, for Sunday and Monday combined, I get visits for , divided by continent, 32.8% from North America, 19.4% from Europe and 0% from all the other continents (south America, Asia, Australia and Africa). That is 22 sessions from America and 13 from Europe while citing a total, of 67 sessions.

Maybe my maths are a bit rusty but I think somehing is not right here.

 

 

 
Top <<<     Mon -- Tue -- Wed -- Thu -- Fri -- Sat -- Sun     >>> Bottom
Day view

MM-lxxv     Wednesday

 

2000-03-15

 

 

While it wasn't freezing (1°C) I did have some ice on the windscreen. The result of a clear night and that type of weather lasted till about midday. Real spring sunshine. The afternoon was gray but still dry.

The day started proficient with two crashes. These were caused by a new screen saver I had installed. Scratch that off my disk and revert back to a standard Flying objects (with a bitmap of my Owl as object :-).

I did search trough OLI for occurrences of the classification codes that must be handled by the HCC system. So now OLI is covered. I also do a quick check for the other modules. That I don't find an occurrence doesn't mean there isn't one of course but I leave finding that up to the testers. If they don't find anything wrong then it is OK by me.

As if it is the most natural thing in the world the "specification" for two new files is changed after I made them. So I have to check them in and change them. These are files to store HTTP info to allow TeleSales to work with non-OLI warehouses using XML. The problem is that there are no formal specifications for the whole action, just a few scraps of paper and a few 'try if this works' programs.
So I start to write specifications now, after the programming and testing is done. With documentation following later still. Not an unusual situation when you are working with computers. Theo tries to prevent this but he cannot do all the preliminary steps on his own. The 24h per day rule cuts into his schedule.

I have been struggling to make some graphics. The problem where I get stuck is that the background must be transparent. Well the real problem is that I don't know how to do that with the tools I have (Micrographics PicturePublisher V6 and Correl V4). I can paint transparent but the original background remains. I know it is possible, it's just a question of time before I find out.

 

 

 
Top <<<     Mon -- Tue -- Wed -- Thu -- Fri -- Sat -- Sun     >>> Bottom
Day view

MM-lxxvi     Thuesday

 

2000-03-16

 

 

Almost dry. Well some barely perceptible drizzle in the morning and another drizzly shower at noon was all we got, while the forecast was for lots of rain. We did get the predicted amount of clouds (lots) and sunshine (none). And the temperature remained at about 8°C all day.

The morning was rather uneventful, with some file checking and picking up some reports from Jan. Ronny did most of the work and I did some assistance. The afternoon exploded however.

In OLI the items are maintained via a download procedure. The actual item numbers are composed in OPRS and then passed via and EDI messaging application to OLI. OPRS is a system built by the MSCC (contradictory named Microsoft Competence Centre). New items can be created in OLI as well but the OPRS is leading. The classification codes (four codes per item) are, in the current release, stored in the CD file but they must be maintained by hand. Because the codes are not mandatory for new items and not checked for existing items (unless they are changed manually) the CD file is never actually maintained.
Now, in preparation to the new r7v9 release a new OPRS to OLI procedure is produced and this requires the CD file to be filled. It doesn't provide a method to do so. And that preparation must be done (and tested) before the new release is installed.

For the HCC project I get the codes from OPRS and build the hierarchical structure. Well that is the intention. The building function is not ready, there are not even specs for it, and the data I got from OPRS for testing was totally inadequate and passed via Excel sheets. Now it seems we will get the data in Microsoft Access format. Of course not in the expected format.
Originally there was no haste to create the HCC filling program but now Loek got the idea that we could fill his CD file based on our HCC file. That probably occurred to him because in the original specifications the HCC file would be built based on the item file and the CD file. Why, we could just, simply, reverse that program to fill the CD file, couldn't we? Right?
Yea for some people life is simple.

I struggled with Access to get the data out and transferable to the AS/400. Well I did not see much of the famous Microsoft user friendliness. I ended up exporting to an Excel sheet and from there saving in a comma-delimited text file (the delimiters are actually semicolons). This could then be transferred to the AS400 where I build a program to parse the text lines and file the HCC file with it. And then another program to fill up the CD file based on the HCC codes. It is all untested yet of course but hey they wanted the program only immediately, not yesterday :-).

Finally I managed to set up Eudora light to retrieve the mail from my ISP account. Whenever I tried using Netscape I never got the mail, only some error message, and I lost the settings for my regular mail account. And using Outlook Express (Win98) I never got mail either but without error messages.

And, to round off a busy day I dug into Correl V4 to find a way to make a transparent background. But I never found out how to do it.

 

 

 
Top <<<     Mon -- Tue -- Wed -- Thu -- Fri -- Sat -- Sun     >>> Bottom
Day view

MM-lxxvii     Friday

 

2000-03-17

 

 

With the temperature fixed (at 8°C) and the clouds locked in place (but no rain) it should be nice to sit behind a screen.

Jan will be away for a few weeks (of to America and then on holiday) so he has been testing a few days now. Finding a lot of bugs that he should have found months ago. Of course they must be solved instantly.

Our intention was to collect the problems and send out a patch each two or three months. Up to patch-4 we have pretty well kept up that system. Patch-4 however was sent out in January, hurriedly and without testing. The result is that we are now fixing bugs and sending out patches at an uncomfortable rate, we are at 08 now. The fixes must be done in a hurry and as everybody knows testing suffers most from hurying. Of course this starts to cause irritation and doubt in the opcos.
All because someone doesn't like testing.
Of course the lack of documentation and specifications is not helping us a lot. Typically a project that has been running wild and, since Peter passed it on, without anyone in the driving seat.

Normally a comma delimited text file has a comma (or a semicolon) at the end of each field. The data I got from OPRS to fill up the HCC file starts off like that but somewhere the last semicolon is missing. The parsing program thus doesn't find the end of the last field and dies (with an index out of range error). The first records are all as expected so while testing everything looked right. I had to step trough all the correct records with the debugger and even then I only found out after several tries. I was looking for a programming error, never suspecting the data (from production) to be corrupt.

Well I don't need Linux or Win2000 to know we are living in interesting times. I just have to look, at the problem reports on (and spilling off) my desk.

 

 

 
Top <<<     Mon -- Tue -- Wed -- Thu -- Fri -- Sat -- Sun     >>> Bottom
Day view

MM-lxxviii     Saturday

 

2000-03-18

 

 

The morning was greeted with a few showers, real rain iso the now common drizzle. This tapered off to a slow drizzle and finally stopped around 10h00. It remained dry but gray and windy the rest of the day.

Shopping. I don't need extra computing stuff so I brought nothing back. Except some tooth paste and a load of wine bottles.

I started to sort and process the stack of papers that is on top of my scanner. Well, that top is flat so it is bound to accumulate paper :-). Checking off papers against files I started to get confused. The problem is that I got lots of duplicate files left over after the last cleanup action. Duplicates are of course good as a backup method because everything is on my hard drive, a Jaz disk and (several) Zip disks. But for backup purposes the duplicates should be the same and they are not.
Well I ended up comparing and unifying files.

 

 

 
Top <<<     Mon -- Tue -- Wed -- Thu -- Fri -- Sat -- Sun     >>> Bottom
Day view

MM-lxxix     Sunday

 

2000-03-19

 

 

It's a cool 3°C with a chill wind but dry and it stays that way until about 17h00 when the sun finally breaks trough the gray stuff. Somewhat later we get a unobscured full moon. So frost will be waiting in the morning :-)

Apart from a few hours gardening, pruning firethorn, I spent the rest of the day on the computer, pruning files and directories.

I tried to upload yesterday and this morning but all I got was a busy line. That is not strange in the evening but this morning, at seven? And it is not a defect because at noon I could get my mail.

 

 


-- Links -- Calendar -- Previous week -- Next week --

Svenson © 1999

A day you don't learn something new is a wasted day.