After a rainy night it dries out in the morning. And, disregarding a few spells of light drizzle it remains tat way. He we even had the sun out most of the afternoon.

The day started off innocent enough. I picked up the preparations for the redesign of the validations where I left off on Friday.

Then Ronny called in sick. Well not really sick. He plays in an indoor football (indoor soccer if you will) team. During the match Saturday he landed wrong after a jump, twisted his ankle and his knee. Nothing was broken but the ligaments of his knee were over stretched and lightly torn. As a result he is home with his knee plastered up so he cannot move it. He is not allowed to drive a car, as if he would be able to with his right leg stretched and immobilised. The plaster should come off by Friday so he is grounded at home with nowhere to go for a week at least.
Unluckily for him his wife isn't allowed to drive either since she has been operated on her eyes. She had a strong negative deviation on both eyes due to an inheritable cornea problem requiring both strong lenses and glasses. Now they surgically corrected her lenses and in a few months time she will get a laser correction to complete the whole process. After that she should have 100% normal vision. Now the operation was just done last week and it will be a few weeks before she is allowed to drive a car again. She is picked up for work by a colleague.
Both are dynamic young and inpatient people. Now locked home in the evenings. Oh dear! <g>

But life goes on in the office. At a pace.
It turns out that the Germans want to start using the Catalog application immediately. That is now. They did mention a few problems with some fields being filled from the wrong basis. The base file is the Item Master file (PT) which is owned and maintained by Logistics (OLI) so while I build the XML program I consulted the OLI team for selecting the right fields. I used the sizes (length, width, weight) including packing. Now the Germans don't have the these values filled in their database. they have the length and width (and height) without packing and the weight with packing filled. Don't ask me why and don't ask them why (maybe they are accepting the defaults TIC). Of course this has to be solved instantly, and shipped on the spot.
So that is what I do. .
Took some doing but I finished it by 18:10. It was not really hard work but time consuming and requiring full concentration especially when you know there will be no time for testing (I got three cups of tea and drank them all cold).
All is ready for shipping. Untested of course.

It didn't rain all day and yet the roads were damp in the evening. Dew. The stuff covered my car in the morning and in the evening. All we need now is a little bit of frost. :(

I missed a few warm cups of tea again today.
The problem was that I applied an external module that gets bound in at compile time. It is a technique we don't often use in Venlo but it works well enough and it cuts down on development time. It's best compared to header files in C though it's not exactly the same, for example the modules bound in don't have to be the same language, you can use a COBOL module in an RPG program.
Another, older method for sharing source code is using copy members. We use two types of these. One is the standard IBM type where the extra source is copied in at compile time only. The other is a home brew where the code is actually inserted before compiling, and remains after compilation. The second method is easy because the extra source is always present. In the IBM type the extra source has to be found by the compiler. The, home grown, version control and development management system doesn't cope with the IBM type of members. When everything is moved from the test/development environment to the sysem-test environment all the objects are tested and removed from the T-env, programs without an object are rejected (reasoning is that if there is no object in T the program probably contains compilation errors). IBM type copy members of course don't have an object (they're just code snippets) and thus are not moved from T to S. Of course the compilation in S cancels because the copy member cannot be found.
We almost exclusively use our own type of copy member and I now know why.

Another problem that had me stuck for quite a time was another program that required a journal file. Of course no journal existed in the S-env. I finally managed to create one and get the bloody thing transferred.
The positive thing is that I now know how to create journals and journal receivers. Small victory.

After wrapping that up, including tests in v7.8 and in v7.9 the whole thing was shipped to Germany.

During our meeting about the XML data encoding last week it was decided that (yeah you know by the powers that be and all such rhetoric) someone should go looking for the availability of a standard XML-parser from IBM. That was after Robby made it clear that on the PC side everybody used the Microsoft XML parser (of course PC and Windows are synonyms for some :). From which Sef and Louis concluded that, on the AS/400, there must exist a standard parser as well. Right. They didn't imply that I had to look for it. I am being out sourced from RealSoftware so I assumed they wanted someone from their own company to do that.
Nope, Theo just asked me to find the IBM parser.
Of course none of the IBM manuals (Blue books nor the Red nooks) carried any information about parsers. I finally found an interface to the parser among a bunch of free but 'unsupported' code (at alphaworks ).
The parser itself is written in C++ and that doesn't interface with RPG or COBOL. The wrapper is as yet not provided by IBM as a licensed product.
But at least it is available. So I can take the next step.

 

Winter is approaching with night temperatures getting below 2C so I picked added a blanket to my bed. Normally all trough the summer I sleep without blanket, just a sheet is enough. And I guess I'd better drop some defrosting spray in the car.

Maybe interesting for Moshe, computers and heat generation.
I had both Miona and Oswal, my two computer on the attic, off all day and by 21:00 the temperature was 12.3C inside and 5.6C outside. Two hours later I had 16.5C inside and 3.7C outside. The attic is quite large (10x6 m and 6m high in the top ridge) and badly insulated and there is no heating. So 2 computers (2x AMD K6 300/350 MHz) do put out a considerable heat.

Again no frost but lots of dew turning into mist and giving way to a beautifully sunny morning. The afternoon was covered up in clouds again with a bit of drizzle to boot. 3C.

I went three paces forward while looking up XML and UTF-8 related stuff. Then I looked around and noticed I was going forward but with my nose pointing in the wrong direction. So I back-track, sidestep and somersault may way back into the right direction. XML is tough (ask Dan ) but UTF-8 and other Unicode related things are really tiring.

We are a bit lucky that Ronny isn't at work now. Just as I had started on the redesigned Validations program Theo comes up with a few extra points to take into account. Nothing dramatic but important enough.
The old situation was that we build the order entry program with all the necessary validations, errors and warnings, included. Along with a few override methods. If an opco wanted some different tests the either had to convince us (tough luck) or make the modification themselves (usually). Of course local versions were under their responsibility.
The current situation in TeleSales is a bit different because the opco doesn't have the sources nor the development tools (ex compiler) so there simply cannot be local versions. This means that we had to allow them to enable or disable each and every validation and we had to provide a bunch of 'local entry' points (calls to dummy programs on the AS/400) where they could add local code. Now the Germans complain that it is easy to switch of validations that actually should never be switched off resulting in 'illegal orders'. (Like you complaining to the manufacturer because your car doesn't stop when you don't push the brakes).
The problem is of course that they can turn off a validation, push an order trough, and quickly turn on the validation. And then complain that the order wasn't validated. (Don't laugh, they do just that.)
Of course we must prevent that.
Well we cannot prevent it but we can log the executed validations so they cannot sneak up on us. And just maybe I add some extra logging of my own.
Of course all that is a change in the design and if Ronny had been here half of the changes would already have been implemented. So he would strongly (and rightly) resist the change.

No frost, only dew that retreats when a beautiful sunny morning arrives, 4C, With clouds and drizzle moving in around noon. It feels just as familiar as it sounds.

A slow day while I work on the redesign for the validations. Doing both the documentation (which is also the specification document) and the coding at the same time.

Tomorrow I got a WebSpeed course, the hands-on type, in order to be able to take over the TOS2000 programs.
Could be fun, could be trouble.
TOS2000 is a Web ordering application that connects to our database. Most of the coding is Progress, which I don't know at all, with a fair bit of Java script, which I know only slightly better, thrown in. The interface is in dynamically built HTML and that is the part that I know more or less. Koen is also following the course. He knows Progress more or less but he neither knows Java script nor HTML.

 

I need a wireless LAN connection.
That should go from a 100Mbps small LAN (12 nodes) to a single desktop PC. The problem is that that PC is on the other side of the street. Just about 20 to 25m away but you can't hang cable over the street. File sharing at the moment is done with Iomega JAZ disks (GigaByte-Sneakernet :) while a standard V90 modem is used for internet and mail. Not the best working environment but not a problem up to recently because the PC was not used intensively. Times are changing though and they want to start using it more.
I checked with the few (more or less) local shops that have a web site and a few local web only shops. Those that have wireless LAN products only carry the 3Com stuff, Airconnect. I assume that is good quality stuff but a wee bit expensive. Just the access point and one PCI card costs over $2000. That is about double what the budget allows for. That budget was inspired by some magazine adds but the magazine was an American thing so the products and prices are irrelevant here.

Very stable weather. Dew turning into sun turning into overcast turning into rain. Only my car was frosted over with just 1C on the thermometer in the garden that is not uncommon. My car was just lightly frosted but half an hour later Suzan's car was completely covered up.

The Progress WebSpeed course progressed (hi) smoothly. We got the whole rundown of setting up a (development) environment. This is a hands-on course so later we had to do it ourselves. Few complications and no pressure on so things went well enough. The old synapses still have some flexibility left in them.
Of course normally we shouldn't have to deal with anything of what we learned today as this is/should be Wilbert's job. Of course when things go wrong and Wilbert isn't available we have to know the routine. And of course this will not happen before we have actually forgotten the whole lot.
Next weeks course is more about actual application development.

That course took the whole day. Well not really, just from 9:00 to 17:00 but in the remaining hours I didn't achieve more than handling mail and rebooting my PC four times. Yep, that is Win95 in action :)

 

I know I seldom post mail, this must be the second or third time, but it may be important (for my pages)

Dave Wootten (bonacker@swbell.net) wrote:

Why is it every time I call up your web page it is white with a yellow border; I start to read your column, then it turns a complete blue (blank), then back to white with yellow again? It's almost annoying, as if you are reading a newspaper and someone yanks it out of your hand!

What's the deal, anyway?

That is strange behaviour indeed.
I use a blue background colour but overlay that with a background image that has the yellow to white gradient. ( <body background="" bgcolor="#00yellow.pngccff"> ).
I use quite some dividing lines and tables with wide borders. In Netscape and Opera any colour attribute given to a line is ignored but the background colour shows trough. The table borders also show the background colour. And I sort of like that. Especially the lines get a sort of 3D, slit like appearance.
The background image is initially only applied to the visible part of the page. Scrolling down should paint the background image over the newly visible area. What seems to be happening in your browser is that the background image is handled incorrectly.
Of course I don't know what browser or what system you are using. It could be a browser problem (old version? odd browser?) or a system problem (low on resources? slow processor?)

The reason why I use this scheme is just cosmetic and I can easily change that if it is annoying. So if anyone else has a similar problem I will put in another background colour.

There seems to be no definitive rule about handling colours, each browser does his own thing and to me it isn't clear which browser is right here: ((

• MSIE
  table borders are always grey, the cell-padding uses the background colour of the table (if defined) or the underlying element (either the page itself or another table),
• Netscape
  table borders always use the page background, the padding uses the background of the underlying element while the cells in the table use the table-background colour (if defined) or the underlying element
• Opera
  table borders and cell padding and cells use the table-background colour (if defined) or the underlying element.

Remembering you can nest tables (table in a cell of another table) and that you can also define background colours for each cell in a table, you notice that things can become uncomfortable quickly. ))

 

Virtual war : http://www.theregister.co.uk/content/1/14798.html

Axiom : something so visible that you don't need to see it.
Budget : a method of worrying about money before you have it
Economy : everything the spouse wants to buy
Intelligence : yes honey.

 

Just a little change in the weather. The morning is dewy and cool (3C) without frost bur it takes longer for the sun to come trough and the afternoon is grey and cloudy but dry. So that is change to come up wit almost the same weather.

Shopping. The usual stuff, apples, some wine, cheese, that sort of stuff. Nothing computerish.
I did point Masset, my usual computer shop, to ELSA. They make modems display cards and screens. They also produce some wireless LAN adapters. Considerably cheaper than 3Com. I have seen positive reviews of their display cards and although that doesn't guarantee quality in their other products it does point to some quality.

I added a surprise hors d'euvre to lunch, brining in a box of oysters. Just 12 for the three of us but then it's just an opener and my father isn't much of a fan for them. Opened an served with a glass of dry Sherry they make a good opening.

And I treated the day, after doing the dishes as a rest-day. Again.

Yesterday dry, today wet. Sums it up more or less. We had rain in the morning and in the afternoon with a few short outbreaks or torrential sunshine.

Warm for running at 8C but enough wind and drizzle for cooling.

The JAZ drive in Timareen, my fathers PC, just died while my brother was making some backups to it. Rather inconvenient. Oh, there is still enough disk space on that box but I used to make backups to a Jaz disk and copy it than on Miona. I intend to get Timareen on the network (a bit difficult to get wire from there to the attic) so that the JAZ wouldn't be needed anymore. That is a long term plan though so I either have to change plans or get a new JAZ drive.
The failure is a bit strange. My brothers disk is giving problems on the other, external JAZ drive while my disks work normally on it. Another disk, from my father, doesn't work on his drive but works normal on the portable one. So It looks like we have a bad disk and a bad drive.
The result is that started testing all the disks on my drive (found no other problems) and I ended up eating my day with shuffling files around.

I got a remark from the ever vigilant Warlock on the problem Dave Wootten alerted me of (Friday). He sees the same behaviour. Now I did some testing and I didn't have a problem at all. Ok on the old 386 with 16MB running OS/2 I do notice a performance problem but I kind of expected that.
This needs some more testing.