My users showed me the weirdest problem today. In my application is a select-box that allows users to select the classification. When the classification is changed, the div containing the icons for the actions (and the classification-select-box itself) is updated to reflect the new classification. So i use an observe_field to watch the select-box, and trigger some javascript to perform the magic. In Firefox this works perfectly. But my users tend to use IE8, for some obscure reason, and switching to FF or Chrome is ... difficult (What other browser?). The observe_field is written as follows: [ruby] <%= observe_field "mention_thm_classification", :frequency => 0.5, :update => "mention-actions", :with => "classification_id", :before => "Element.show('spinner_cl')", :loaded => "Element.hide('spinner_cl')", :complete => "Element.hide('spinner_cl')", :url => { :action => "set_classification", :id => mention } %> [/ruby] Pretty straightforward. The observe_field is contained inside the div that is updated. When in IE8 a new value is chosen, i get the correct value in my method first, and then i get a whole bunch of classification_id="null". I am guessing what happens is that when the html is updated, somehow also the select-box is emptied, and thus changed, and thus a new value is triggered. So i replace the html again, and the select-box is emptied again, and an endless loop is born. So, the easy solution should be: do not do the action of the observe_field if the value of the select-box is null, and that is actually really easy: [ruby highlight="3"] <%= observe_field "mention_thm_classification", :frequency => 0.5, :update => "mention-actions", :with => "classification_id", :condition => "value != null", :before => "Element.show('spinner_cl')", :loaded => "Element.hide('spinner_cl')", :complete => "Element.hide('spinner_cl')", :url => { :action => "set_classification", :id => mention } %> [/ruby] We add a :condition that checks just that! It still feels i am doing something wrong, that i should not have to do this, so i would be happy to hear any way to improve this.

In my Rails application i use a generic application.rhtml for all my pages. Amongst others, this code contains some jQuery code to manipulate my page [javascript] var $j = jQuery.noConflict() $j(document).ready(function() { // remove all alt-tags $j("[alt]").removeAttr("alt"); // validate all forms $j("form").validate({ errorPlacement: function(error, element) { error.appendTo( element.parent() ); }}); }); [/javascript] The things i do in my document ready callback is simple: remove the alt attribute everywhere, so my tooltips will work on my links containing images, and if there is a form on the page i validate it. Now for one form, i need to add special code, so that when a user fills something in a field, a select-box looses some options. The ideal place to do this, is in the document ready. But not for all pages. How do i solve this in some elegant way, and not use a different layout for that page. I choose to use yield and content_for. Like this. Add the following line : [javascript] var $j = jQuery.noConflict() $j(document).ready(function() { // remove all alt-tags $j("[alt]").removeAttr("alt"); <%= yield :script %> // validate all forms $j("form").validate({ errorPlacement: function(error, element) { error.appendTo( element.parent() ); }}); }); [/javascript] and in my view i write: [ruby] <% content_for :script do %> var allOptions = $j('#mention_thm_classification option').clone(); $j('#mention_thm_reference_local').change(function() { var filter_txt = '.ref-'; if (!$j(this).val()) { filter_txt = filter_txt + 'empty'; } else { filter_txt = filter_txt + 'filled'; }; $j('#mention_thm_classification').html(allOptions.filter(filter_txt)); }); <% end %> [/ruby] So what i actually do is, check whether the value of a field changes. When it contains something a select-list with id mention_thm_classification is changed, and only the options with the correct class are kept. For completeness, my select is built as follows: [html] <select name="mention[thm_classification]" id="mention_thm_classification" > <option value='1' style='background-color: #b7f9e2' class="ref-empty " >Green</option> <option value='2' style='background-color: #ffdebb' class="ref-empty " >Orange</option> <option value='3' style='background-color: #ffbbc6' class="ref-empty ref-filled" selected="selected">Red</option></select> [/html] So i have described a nice way to add extra code to my document ready function without needing a special layout-page. And experienced for the umpteenth time how great jQuery is.

I am using the jQuery validate plugin which is awesome for straightforward form validation. What i especially like is that it offers the user a better user experience, because there is no need for the round-trip to the server. I still need to do the validation on the server-side (rails) as well. I guess there should be a way to extract the validations from the model into the view transparently, but for now i just add a class "required" myself. Now i did bump into a seemingly difficult problem, when the user wanted my to validate to having either one of both fields filled in. Luckily google came up with a beautiful answer quick enough! I needed to adapt it somewhat, because my form-styling places each label-field combination inside a div, so i search for the second surrounding 'div' (containing the group of fields for which only one field is required). [javascript] jQuery.validator.addMethod('required_group', function(val, el) { var $module = $j(el).parent('div').parent('div'); return $module.find('.required_group:filled').length; }, 'Please fill out at least one of these fields'); [/javascript] How i love the jQuery and Rails community. Awesome!

In my Rails application, at a certain stage, it was necessary that my user performs an operation in another web-application (a GIS application) on the same database. I am using the following piece of code to make sure that i open a url in a new window, and that it should get the focus (jump to the front). [javascript] <% javascript_tag do -%> function newWindow(newContent, windowName) { var new_window; var new_window = window.open(newContent, windowName); if (window.focus) new_window.focus(); } window.onload=function() { newWindow('<%= link_to_location @mention, true %>', '<%= @gisapp_window_name %>'); setTimeout(jump_to_select_tab_url, 2000); } <% end -%> [/javascript] All is well when the window is first opened, but if the window was already opened, still it fails. And of course i do all testing against firefox 3.5, although my user wants to use IE instead. So i try in IE and it works. Huh? So it has something to do with Firefox 3.5. And behold, as explained here, firefox is so smart it will block all javascript manipulation of windows, which is of course just what we need in this case. So inside firefox, i need to go to Options->Content->Javascript and make sure that i can run javascript that will manipulate opening and closing of windows. Done!

I am using Rails 2.3.4 at the moment, and as is well known, rails uses Prototype as default javascript library. For most purposes this suits me well. But what i really like about jQuery is the wealth of plugins readily available. My solution was not something like jRails, but i use them side by side. This might not be the best solution :) First of, you have to include the correct javascript files. Then, at the top of my application.rhtml, i write this: [javascript] var $j = jQuery.noConflict() $j(document).ready(function() { /* insert some jquery magic here */ ... }); [/javascript] This ensures that i can use the jQuery library nicely alongside the Prototype one. Just replace $ by $j for all jQuery calls. For most jQuery things this is more than enough. But then I needed form validation, and needed to be able to call some jQuery from inside a link_to_function block. It took me a while to find, but actually (isn't it always) it is dead-simple. [ruby] def add_document_link(divname, parentdoc ) link_to_function get_icon('adddocument'), { :title => t('adddocument')} do |page| new_doc = Document.new new_doc.doc_ident_p = parentdoc.doc_ident page.select('.add-items-link').each { |b| b.hide } page.insert_html :bottom, divname, :partial => 'documents/add_document', :object => new_doc, :locals => { :divname => "add-items-link"} page << "var $j = jQuery.noConflict(); $j("form").validate();" end end [/ruby] Easy as pie :) I love rails :)

I am developing Ruby on Rails on Windows platforms mostly. But using Rspec and cucumber on windows has a very strange side-effect: all a's are missing as can be seen from the following screenshot: Luckily, after some very extensive googling, i found a single blogpost with a fix! Apparently it has something to do with UTF-8 encoding, and the simple solution is that you need to change the encoding of the current command prompt. This can be achieved via a simple call before you start: [sourcecode] chcp 1252 [/sourcecode] The aforementioned post then proposes to adjust the cucumber.bat to not have to type this every time. This is all good for cucumber, but not for rspec, and anyhow, every new update of the cucumber i would need to apply this fix again. I was thinking that it might be possible to set the default codepage, which is 850 on my machine, to 1252 permanently. As this blogpost mentions, there are two different ways to achieve the wanted result.

Change the codepage system-wide

This can be done in the registry. [sourcecode] [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Nls\CodePage] OEMCP=1252 [/sourcecode] But one commenter notes this is not without risk.

Only for command prompt

An alternative way is to put make sure that each time a console is opened (cmd.exe) the codepage is set automatically. [sourcecode] [HKEY_LOCAL_MACHINE\Software\Microsoft\Command Processor] Autorun=chcp 1252 [/sourcecode] This will only work for console windows in which you run cmd.exe, which is just what i needed.

...or software patent troubles again

The weirdest news i read yesterday: I4i is not out to destroy Microsoft? In the article is described that some small company, i4i, has apparently won a case against Microsoft, claiming the Microsoft Word violates a patent they hold. More specifically: patent5787449. The Texas judge ruled that Microsoft has to pay a fee/fine of 200 million dollar, AND that Word versions 2003, 2007 and future versions can no longer be sold. And the problem only arises when clients or customers use a feature called "custom XML". What? Let's rewind here for a bit. What does that patent actually say? Let me summarize it for you. It describes two things: 1) it describes a system where the content is seperated from the actual representation, and 2) using a dialect of SGML. Now note: this patent was filed in 1994, before XML even existed (only since 1996), and the patent was finally issued in 1998. I tried to read the patent-document as some kind of specification, but i am not even sure what it is actually protecting. An algorithm? The idea? Because the idea of seperating content and representation is a common idiom. We all do it. Websites do it. We use css all the time. We use templates in Word ... The idea of using metacodes (like Title, Chapter, ...) and then looking up their definition the replace in the actual document (this is the example from the patent). So it must have something to do with the specific implementation: they use an XML file. Now comes the really scary part. XML is really powerful, and Word is using a set of XML files, packaged in one zipped archive as their file format since Word 2003. This is really awesome. This is more open than any binary format. Now anyone who can read can actually create or process Word-files. Microsoft calls it Open Office XML or OOXML, but that doesn't really mean the standard is really open. Only Microsoft can amend the standard, and the other big standard, OpenDocument (ODF), is the fileformat originally used in OpenOffice.org (you see the deliberate confusion in the naming?), is totally different, not supported. A missed opportunity. But that is another discussion all together. Now, i4i has no problem that Microsoft uses an XML file, just as long as the format is fixed. Meaning: you specify which tags you can read, are defined (for instance using DTD), and those tags are fixed. But Word allows you to add Custom XML, you can add your own fields into any Office document. And i4i claims that is the problem. So now i am lost. This has nothing to do with seperating content and presentation (that was what the patent was about, right?). It could have something to do with how the extra tags are added, the custom XML. If they would use a lookup-table, to identify the custom tags, then they use a technique that is described in that patent. But hey, wait, does that mean you can not use custom xml tags anymore? I have used tons and tons of xml, with no predefined DTD, so that the file format is extensible, and users can add their own extensions, and keeping the definition of the tags seperate. In my personal opinion, and i think i am not alone, Microsoft is falsely convicted. There is no technological ground this patent can hold. This all comes back to the old software patent discussion. What is patentable and what is not. It seems so unnatural to patent software, algorithms, they are the building blocks upon which we learn and build the actual stuff. And the actual stuff (software/websites) is most of the times so general on the one hand, but very specific to the customer, i don't see how these could be patented. I read that the way a website operates can be patented in the USA. Back to the patent '449. In the press release found on i4i's website, they claim that Microsoft infringes claims 14,18 and 20 from the patent. Those claims refer to the ability of using a lookup-table, being able to generate it, being able to read and replace the values, and being able to use more than one. That is so general! They could sue anyone who is using xml and some kind of look-up mechanism. I think the patent should just be declared invalid. But i guess that is easier said than done. Speaking about the seperation between content and representation, speaking of metacodes, reminded me of Donald Knuth, inventor of TeX, and author of the greatest series of books of all time: The Art of Computer Programming. So let me point you to a letter by that same man, claiming mathematical ideas or algorithms should not be patented. Final thoughts: i4i LP is a company that guards the patents from i4i. So this could be some kind of patent troll? Secondly, they choose Texas, where a large number of these kind of cases have been won, because of their technological inaptitude. I am looking forward to Microsoft's next step. What are your thoughts?

PL/SQL package: lessons learned

Working with Oracle Spatial a lot, I found a great source of information in the Spatial DB Advisor, which besides a lot of interesting articles, also offers his source-code and PL/SQL packages for free. The site is not very user-friendly (trouble of seeing the threes through the wood), as finding the free packages is something that i only succeed in coming from Google :) A shortcut: you can find them here. But a word of warning: if you try to install this package in an existing schema, it will delete all indexes and tables. This is something i had to discover the hard way (painful i might add). So i have adapted my version of create_test_data.sql, where at the top of the file all indexes and tables of the user are dropped. Nevertheless, extremely useful package. I will list my favourites (for now):

• isCompound: checks whether a geometry contains any circular/arc elements
• ConvertGeometry: converts any special elements - circulararcs, rectangles, circles - in a geometry to vertex to vertex linestrings
• to_2d: converts a sdo_geometry to 2d. Very useful for us, as geoserver can't draw 2d, and we only use 2,5d anyway. Meaning that we always use a topview, and the z is the elevation level of the ground-level. Or for pipelines: below ground :)
• sdo_mbr: function to determine the minimum bounding rectangle! i will use this to adapt my script to fill user_sdo_geom_metadata. Cool :)
• functions to investigate the data of sdo_geometry (number of rings, number of vertexes, number of coordinates, the coordinates itself, ...)
• functions to investigate the meta-data
• functions to manipulate your sdo-geometry: scale, rotate, affine transformations, move, adding and removing points, ...
• ... and lots more ... In short: awesome!! So after getting a terrifying scare, now rebuilding the databases (and adapted the script: won't happen to me again). I mailed the original creator to add a note of warning on his site too. Might help other absent-minded developers :) I also had another smaller problem, because the install-script overruled my ORACLE_HOME and PATH definition, which made sure SQL*PLUS was not found. But that was easily cured. I am very grateful that he puts up this package for free. And i learned another great lesson today: not to get too enthusiastic before i am sure that it all works ;)

I have been trying the Windows 7 beta, and i must admit it is pretty awesome. What is less awesome is the fact that now the RC is released, and i have to reinstall all my Windows 7 machines (three --i admit i was a bit too enthusiastic). And if i do that, once the final version is released, i have to re-install them again. For my standard day to day use not too bad, but not for my development machine. So i am back to Windows XP now. With a great new look: Zune XP Theme. Nice :) Working with a lot of open source products lately, so it would be nice to use an open source windows alternative. For the moment it would still be very hard for me to just not work on Windows. I still have to develop for .NET and MicroStation and AutoCAD, all only working on Windows platform. So I hope that ReactOSwill become more mature soon. It is really improving fast lately, but still not completely ready to replace Windows. I will have a test-drive real soon. What are the other options? Run Linux and do my Windows work inside a VM? Buy a Mac and do the same? Any suggestions?

When storing spatial data into Oracle, there are a few steps one needs to complete (as shown here too):

• insert the data (obviously)
• update the USER_SDO_GEOM_METADATA table. This table specifies for each column the bounding box and the SRID (coordinate system).
• create a spatial index I have created a script to this automatically for me, once all tables are filled with their data (e.g. after import or after using FME to import your data). [sourcecode language="sql"] set serverout on DECLARE schema_orig varchar2(100) := upper('&user_orig'); CURSOR ctab is select TABLE_NAME from sys.dba_tables where owner = schema_orig; tn sys.dba_tab_columns.TABLE_NAME%TYPE ; CURSOR ctabcol is select column_name,DATA_TYPE from sys.dba_tab_columns where owner = schema_orig and table_name = tn ; collist varchar2(2000) ; query varchar2(2000) ; column_name varchar2(200); has_geometric_column boolean; col_count number; BEGIN -- dbms_output.enable(1000000); FOR ctabrec IN ctab LOOP tn := ctabrec.table_name ; has_geometric_column := false; col_count := 0; for ctabcolrec IN ctabcol LOOP if ctabcolrec.data_type = 'SDO_GEOMETRY' then has_geometric_column := true; col_count := col_count + 1; column_name := ctabcolrec.column_name; -- insert data into user_sdo_geom query := 'INSERT INTO USER_SDO_GEOM_METADATA (TABLE_NAME, COLUMN_NAME, DIMINFO, SRID) VALUES ('''||tn|| ''' , '''|| COLUMN_NAME ||''', MDSYS.SDO_DIM_ARRAY( MDSYS.SDO_DIM_ELEMENT(''X'', 0, 10000000, 0.000005),MDSYS.SDO_DIM_ELEMENT(''Y'', 0, 10000000, 0.000005), MDSYS.SDO_DIM_ELEMENT(''Z'', 0,10000000, 0.000005)), &srid)'; dbms_output.put_line(query); BEGIN execute immediate query; EXCEPTION WHEN OTHERS THEN dbms_output.put_line('**** failed to insert row into user_sdo_geom_metadata for '||tn||'('||column_name||')'); --rollback; ignore any errors! there will be existing columns END; -- create spatial index query := 'CREATE INDEX '||tn||'_'||to_char(col_count)||'_SX ON '||tn||'('|| COLUMN_NAME ||') INDEXTYPE IS MDSYS.SPATIAL_INDEX'; dbms_output.put_line(query); BEGIN execute immediate query; EXCEPTION WHEN OTHERS THEN dbms_output.put_line('**** failed to create spatial index for '||tn||'('||column_name||')'); --rollback; ignore any errors END; end if; END LOOP; END LOOP; commit; END; / [/sourcecode] The script has two parameters: the schema from where to scan all tables, and the srid (coordinate system) which needs to be filled in. Any thoughts on this? suggestions?

I am now working in Ruby on Rails for about a month. I have done a few smaller things before, investigating, trying out; but now I am really building a full application site in Rails! It feels great :) But of course, just starting out in such a new environment, both ruby and rails, I encountered a lot of initial problems. One of them is showing international, accented characters, like é, è, ê ... from the database. Nor Rails nor irb could show the correct characters. I really needed to solve this problem, so this started an investigative journey.

The symptoms

When displaying a table of results from Rails in any browser, all accented characters were shown as white question marks inside a black rhombus (diamond?). When i tested this with irb i also got very weird results. When using TOAD the results of a query are displayed correctly. But maybe TOAD just does it very clever? If i use the GUI version of SQL*Plus (sqlplusw) the results are also shown correctly. But when i use the console (command-line) version of sqlplus, the characters were also distorted. So maybe the clue could be found there.

Oracle NLS-LANG

No matter what character-set the Oracle database uses to store the data (e.g. UTF-8, UTF-16) the data is always converted to the clients character set. This is truly an amazing feature. This depends on a setting called NLS_LANG. Not sure wat it quite stands for. NLS_LANG is built up as follows:NLS_LANG=language_territory.charsetThe NLS_LANG property is set by default in the registry to the correct Windows codepage.NLS_LANG="AMERICAN_AMERICA.WE8MSWIN1252"But why doesn't it work inside the command line console? Apparently because it used a different codepage. If you type chcp (requesting the codepage) at the DOS prompt, it would normally return 437 (it did on my machine). So you would need to enter`

set NLS_LANG=american_america.US8PC437 require 'oci8' OCI8.new('user','pw','db').exec('select * from emp') do |r| puts r.join('|'); end `in your irb and then all results would be displayed correctly. The crucial line being the correct setting of NLS_LANG. Wow! I got my results correctly in ruby! Now i was in the assumption that Rails would be a piece of cake, but that was wrong.

Fixing Rails

The easy idea would be to set NLS_LANG in Rails correct, before the oci8-library is required. My first approach was to set the NLS_LANG in the first line of the environment.rb with the following line:ENV["NLS_LANG"] = "AMERICAN_AMERICA.WE8MSWIN1252"But this didn't work. I am using NetBeans 6.5, and it took me a while to realise that if I edited a file to contain special characters (fixed text, e.g. on a menu) it would work. NetBeans (or Rails for that matter) standard works with UTF-8. So all files are encoded in that way. The easy solution would be to useENV["NLS_LANG"] = "AMERICAN_AMERICA.AL32UTF8"But that didn't work. I tried the alternative AMERICAN_AMERICA.UTF8 but that didn't work either. I am just guessing here, but i think somehow the NLS_LANG setting didn't get picked up in the Rails environment. It kept using the registry setting. So I tried to turn it around: make Rails use the windows codepage instead of utf-8. This took several steps:

• change the default setting inside NetBeans to use the correct code-page instead of UTF-8. This in fact only affects the format of the files that are saved, but it is nevertheless important that all files being served are in the same format. Also convert all previously edited files from UTF-8 to standard ASCII. I used an editor to do that.

• added the following to application_controller.rb:` before_filter :headers_iso

  def headers_iso

  make sure the charset matches the default Oracle NLS setting

  headers["content-type"]= "text/html; charset=windows-1252" end `

• added the correct META-tag to application.rhtml in the HEAD-section:``Note: this meta-tag is not really needed, it has no real effect. And that finally worked! :)

I have this machine at home, running XP, and all of a sudden it is getting incredibly slow at startup. I installed Ad-Aware to make sure i don't have any hidden spamware running. It found nothing. So then i tried to find ways to make XP run or boot faster again. I know that the really failsafe way is re-installing my XP but i don't really have the guts to do that, reinstall all drivers and programs. So i tried my friend, Google. I found several articles allowing me to tune my xp-performance and startup time, i applied the ones i found sensible, the ones that were mentioned accross different pages. In short i think the best things you can do is uninstall unneeded programs, make sure that unwanted programs in the background can't start (msconfig). Bootvis seems interesting too, not entirely sure if it helped a lot. Defrag of course! :) Making sure your pagefile can't get defragged by making it large enough, preferable on a separate disk from your boot-drive. These changes seemed to improve my boot, although it is no where near a quick 20 seconds now :) It still takes a few minutes to get my computer completely ready. But suddenly i also lost my network connection. I was puzzled, did i stop one service too many? :) Everything seemed to work fine, but my network-card did not look for an IP address (DHCP). I tried to start my Firewall and i got an error stating the WSAStartup had failed to start, or was not started. WSAStartup? What is that? Looking on the internet i found a small tool, fixing WinSock problems, possibly related to de-installing programs. You never know. And i had no other clues to follow anymore. The tool is called WinsockXPFix. It supposedly could also backup my registry, but that failed in my case. So i just started it, in blind faith, knowing i had to re-install anyway if it wouldn't work. But guess what: i did! The program ran, rebooted automatically, and everything was back in working order! Windows XP, at the time, was promoted for its boot performance, it's self-tuning capabilities, yet over time it only becomes slower and slower. How does Vista handles this? Or Windows 7? Or is it inherent to Windows OS, and should we look at other alternatives. Sometimes i hate how much is done behind the scenes, how much is added. Each vendor or software installs its own version of "Automatic Updates", slowing down my system. Should I look at Mac OS X? Or Linux? Yet Linux still feels to "hardcore" for me, but maybe i am wrong. Should we look at different open source alternatives (in order of viability):

• Visopsys? this seems like a single developer's playground. Not quite sure what kind of future this holds. But what a great effort to complete alone, though :)
• Syllable? not quite sure about this one either, either a Linux-based server, or a desktop which is based on Amiga :) That is very nice, i sometimes long for the old times with all the different operating systems out there: QDOS (Sinclair QL), AmigaDOS, Acorn Archimedes, Next. But reviving something like that will never be able to compete with anything commercial. Although i know Mac OS X is heavily based on Next, and rightly so. Next was anyway far ahead of its time. Not to mention they looked nice. I have a soft spot for matte black :)
• SkyOs: this seems promising, seems to be based on BEOS, which was very promising at the time. It has a lot of applications (open source) available, and a concept of a Software Store: a central place to find all software. A bit like AppStore maybe, and sounds like a good idea.
• Linux, it all its different flavors
• Wine allowing to run Windows-application under Linux, makes Linux even more interesting :)
• Hackintosh: running Max OS X on cheap intel computers. Sounds like a viable option too, albeit a little illegal, since Mac OS X is only allowed to run on Apple hardware. That is a shame though. Would be tempted to try that, just to know for sure Mac OS X is my thing before i buy such a machine.
• ReactOS: i have no idea how good this is at the moment, but this sounds like an extremely interesting idea to me. Create an open source, improved, version of Windows XP. You can run all current Windows applications on it. I guess they hope to stay up to date. Looks like a very interesting option. So i am also not quite sure what the impact should be, changing OS. I guess application-wise it doesn't matter too much nowadays. There is Open Office, AbiWord, .. which allow you to most of the day to day jobs. Not sure if i could miss PhotoShop though. There is GIMP, and GIMPShop should make switching a lot easier. And such a new OS could revive my older hardware. Maybe.

It has been the strangest year for me, professionally. I had a good job, where i was hired to do C++ in a Windows CE environment, but where i ended up doing the back-end in SQL Server and C#/.NET. Where i also learned about WPF and WCF and was really impressed. WCF is really easy. Dead-simple. But I really love WPF! It feels like HTML/CSS for Windows UI, except more powerful :) Granted, it took me a while to really get to grips with it, it is not as easy as WinForms is. Although it is not complicated, it's just a different way of thinking. Then i got another offer, from a company where i worked before, long ago, as a youngster, before i became self-employed. I was not really looking, but it felt welcome. I was reluctant somehow, because the other programmers in the company would be leaving, but they promised that would give me the chance to make a new start. I told them i was not going to be able to do it all alone, and needed support. That would not be a problem, they said. So back to geographical information systems. It sounded awesome! I saw the product they were selling, and we could improve that easily. It was web-based, so that would be my chance to finally get into that. I was looking at all kinds of technologies, thinking of .NET, SilverLight, Ruby and RoR. When i arrived there was one guy, not willing to share any info with me. I had to maintain and keep 5 projects up and running on my own, while i didn't even know them from scratch. Communication, like in with a client, is not really my cup op tea. I think i am good at explaining a program, sometimes i think i would like teaching, but i am not good at giving commercial info. All i could think off: sorry, i am all alone here, so i think you will need to wait a few months before i can start on your project. Which i assumed would not sound good, but it was the truth, so i kept silent. Which i believe, in all honesty, is even worse. I told our so-called commercial guy that we needed to communicate to all our clients they were going to have to be delayed, i needed to be worked in, i needed time, ... But all he did was trying to push me to take care of his pet-project. The other projects he just didn't care about (or at least: didn't seemed to). At the same time, there were also a lot of requests for changes to the software used inside our own company. And i was looking around at the possible future implementations. New technologies. It was a though, hard, and stressy time. Too many balls to keep in the air, and juggling never was my strong point. Now, we are almost 5 months further. I have implemented software in VBA for MicroStation and AutoCad. If they had told me beforehand most of my implementation work would be VBA, i might reconsider :) But let's not get into that. I have written one application to calibrate data along a pipeline, in C# and WPF, and show it graphically. That was awesome. We hired a junior, which I try to coach. Not very good i am afraid, with all the work pressure. But coming January i will have a new boss (project manager/commercial) and two more new collegues shortly after. The ambition of my new boss is simple: become the market leader in what we are doing now. Great! That sounds awesome. The last year i have been looking around at so many new technologies. I have played a bit with ruby. I tried out Ruby on Rails. I want to get my hands dirty. And soon! But we need to soothe our clients first, which means: improving their existing systems, before we continue on and create something new. Has anybody ever heard of WebDev? Well i hope you haven't, but i need to work in it. It should be extremely powerful, but maybe i just don't see that yet. Patience is a virtue.

At the company where i am working now, i need to hire new developers/programmers. Remember, the IT-department of this company is constituted entirely by me alone. Problem is: how to recognize a good developer? Luckily there are some very useful articles available on the net:

• Joel's Guerilla-Guide to Interviewing
• Interview questions for Candidate Software developers (can you see where he got his inspiration from :))
• Write a String Reverser (and use recursion!) It has been a long while for me to interview people. I don't want it centered around a specific technology, i just want to be able to grasp, somehow, whether or not they are good developers. Good developers, which i consider myself to be, will get their head round a new programming language in no time. I feel these interviewing hints (mentioned above) will help me finding the right collegue. On top of that, the environment I am working in is still a bit volatile, we could still choose for any valid technology at the moment (java/j2ee, C#.NET, RoR). I guess it will depend on my new collegues experience (and mine). I am favouring either .NET (because i know it well, because it is very developer friendly) or RoR (new, but very promising). If i get a very good java collegue, i am willing to go down that road too.

At my new job I am back to administering Oracle. The last time I had to do any dba-job, they were still using Oracle 7 :) At my current job, the IT-department can be summed up in one word: me! So, now i have to do everything again. So today, i was trying to get a database installed on a laptop left behind, with no passwords. I was able to solve that, luckily. I had a script to create the database in Oracle 8, had to port it to 9. The biggest change there was: instead of svrmgrl you have to use sqlplus. So i could run the script, but kept getting ORA-01031 error (insufficient privileges). At first: didn't use a sys-account, then lost password of sys-account (set it again using Enterprise Manager), logon using sys-account, but my network user had no ora_dba rights. Logon with a local user with adminstrator rights, put him in the right group, and then it still didn't work ... And at those times, I am so happy there exists this something called internet. I found my solution here. I would never have thought of that. Go figure: a whole day spent trying to create a database. MS SQL Server or MySQL seem so much easier. Industry leader or not, you would suppose they could make equally simple. Or maybe i just haven't found it yet. Don't mention import and export. That is also a pain. But once it is runnning ... :)

Both AutoCad and MicroStation can now be automated/extended using VBA. This is fabulous. In my current working environment, these programs are used extensively, and a lot of software has been written for it. But unfortunately as well, a lot of programmers have come and went again. So i was left with a program to change, which was locked using a password. A VBA password. I am not quite sure why one would do such a thing, one valid reason would be that some user would accidently get into to the code, and change stuff? No idea. Stop reverse engineering? I could get into a whole discussion about protecting knowledge and sharing knowledge. Anyway, i needed to get in the code, and the person was either unreachable or not willing or forgot the password himself. So i looked for ways to break a VBA password. Apparently there is a very simple way, which took my hours to find. In the *.dvb file, look for the tag "DPB=", and replace that with "DPx=" using a hex editor. Load the project, click through all error-messages, open the properties of the project, and set the password again to something you know and remember, and after that you can remove the password protection all together (if you want). A more detailed description can be found here. This method seems to work for Excel, AutoCAD, let me know if it works for other VBA password protected documents or applications (e.g. Word/Access/... ??).

I find the new Firefox 3 beta 5 unbelievable. It is much faster than the older version. But one of the new "features" bothers me a lot. I don't like the new address-bar autocomplete function. The location bar is now very clever, and if you start typing a word, it will look for all similar words and tags in your bookmarks and history. This is all swell if you don't know the exact url where you want to go, but most of the time i do. Then this option becomes counter-intuitive.

And frankly, i don't like how it looks. It is too big. Too much in my face.

Maybe i still need to get used to it.

But there is an alternative. There are two ways in which the behaviour of the urlbar can be configured.

If you type "about:config" in your location bar, you can configure Firefox 3 beta 5. If you then look for the browser.urlbar there are two relevant settings:

• matchOnlyTyped: set this to true, and you get more or less the same behaviour as before, it will only match on url, but it will still sport the new look
• maxRichResults: if you set this to zero, you disable the location bar autocomplete function entirely.

So it is up to you, which you prefer.

Mind you, if privacy is the issue, like embarassing url's popping up while typing an url (who likes to admit an unhealthy Disney addiction), record-keeping in Firefox can now easily be disabled. Just open Extra-\>Options-\>Privacy and unselect the "keep history for at least ..." checkbox.

When looking at examples and samples for the .NET System.Xml.XmlTextReader, i was hoping it would be as easy to use as XmlTextWriter. But i did not find such examples.

XmlTextReader will read node per node, using the method Read, and a programmer will have to evaluate what type of node the current node is. But XmlTextReader has two short-hand methods: ReadStartElement and ReadElementString.

Suppose we have the following xml :

 <root>
   <itemlist>
     <item> 
       <name>Item1</name> 
       <value>110</value>
     </item>
     <item>
       <name>Item2</name>
       <value>234</value>
     </item>
     <item>
       <name>Another Item</name>
       <value>432</value> 
     </item> 
   </itemlist>
 </root>

What we want to do is read one item in a simple way. That is possible as follows:

 private void ReadItem(XmlTextReader reader, out String name, out String value) { reader.ReadStartElement("item"); name = reader.ReadElementString("name") ; value = reader.ReadElementString("value"); reader.ReadEndElement(); //item } 

Unfortunately ReadStartElement will throw an exception if the current element is not the same as the wanted element. So we need a bit more code to iterate smartly over all possibilities.

private void ReadItemList(XmlTextReader reader) 
{ 
    reader.ReadStartElement("itemlist"); 
    reader.Read(); 
    // read next element 
    while (reader.LocalName == "item") 
    { 
        String name; 
        String value; 
        ReadItem(reader, out name, out value) ; 
        // ... do something here with read item ... 
        reader.Read(); 
    } 
    reader.ReadEndElement(); // itemlist 
}

The reader.Read() positions ourself on the next node, allowing some kind of look-ahead, while the ReadEndElement and ReadStartElement can handle this perfectly well. They will either proceed to the next node, or stay on the current node. That's why this code works. For completeness, this does the setup:

XmlTextReader importReader = new XmlTextReader(fileName); 
importReader.ReadStartElement("root"); 
ReadItemList(importReader); 
importReader.ReadEndElement(); 

... where filename points to the file containing the above shown xml.

If you have any questions or comments regarding this example, please let me know.

This allows for very readable and concise code, but only if you know how the xml file you are reading is formed (the order of the elements). I will solve this in the next post.

WCF and MaxStringContentLength on XmlDictionaryReaderQuotas

When sending xml-data over WCF using a string, i bumped into the following error :

The formatter threw an exception while trying to deserialize the message: Error in deserializing body of 
request message for operation 'Test'. The maximum string content length quota (8192) has been exceeded 
while reading XML data. This quota may be increased by changing the MaxStringContentLength property 
on the XmlDictionaryReaderQuotas object used when creating the XML reader.

Right. This sounded like some foreign language where i can't even make out the syllables. I am using WCF with a NetTcpBinding, so after extensive googling i was able to produce the following C# code :

    ushort myport= 27001;
    const Int32 myMaxMessageSize = 6553500;
    string address = String.Format("net.tcp://localhost:9000/myservice/{0}", myport);
    XmlDictionaryReaderQuotas quotas = new XmlDictionaryReaderQuotas();
    quotas.MaxStringContentLength = myMaxMessageSize;

    NetTcpBinding binding = new NetTcpBinding();
    binding.ReaderQuotas = quotas;
    binding.MaxReceivedMessageSize = myMaxMessageSize;
    binding.MaxBufferSize = myMaxMessageSize;

    m_factory = new ChannelFactory<imyservicecontract>(binding);

    m_proxy = m_factory.CreateChannel(new EndpointAddress(address));

Setting the MaxStringContentLength alone doesn't cut it. Once your buffer grows larger than 65K (the default for MaxBufferSize and MaxReceivedMessageSize ) you will run into errors again. So that's why i opted to change those as well.

Hope this helps :)

Well apparently MS SQL Server 2005 does not accept remote connections by default. Luckily i was not the first to bump into this problem, here is a step-by-step explanation to solve that. Kudos to those guys! :)