Illuminaries in the Maldives

MCPD Official Launch

2010-12-03T02:34:00.001+05:00

The Maldives Child Protection Database was officially launched on 24 November 2010 with Maldives Vice-President, Dr Mohamed Waheed Hassan Manik as the chief guest at the ceremony.

Mohamed Naeem, Child Protection Specialist in the UNICEF Maldives Country Office, describes this event as marking "another milestone in the work that we are doing to improve the situation of children in Maldives and in or work to achieving the right to protection for Maldivian children". He also noted that "The good work done by Illuminaries was highlighted by everyone. I personally would like to highlight the extra mile you went to get to this point. So a very special thanks to you both [Tony and Jim]".

Following the launch we have received and been copied into a number of emails from others who have had an involvement in the project.

From Joshua Kakaire - formerly UNICEF Maldives Country Office

"Congratulations Naeem and team, it is more your resilience and relentless efforts that have yielded this historic achievement. Now the real stuff starts; populating it, generating reports and utilising the data for the betterment of children and women in 'Paradise'..."

From Manssor Ali - Special Advisor, UNICEF , New York

"Naeem your hard work and persistence has paid-off. I would like to express my thanks to all of you as well as Illuminaires for their dedication and patience in delivering a user sensitive system that will help in transforming the work of the Government to plan and effectively deliver protection measures to the troubled communities and families in Maldives."

From Abheet Solomon - UNICEF Tanzania Country Office

"Congratulations to you, UNICEF Maldives team(s), Govt Ministries and Illuminaries for making this happen!!! Reminds me of an old African proverb "Ideas are worth a dime dozen, but people who put them into action are priceless"

There aren't very many examples of such integrated multi-sectoral systems, and the launch of this is excellent news."

Report Viewer Control

2010-05-27T19:59:00.009+05:00

I thought it would be useful to blog something about the ReportViewer control which was new to VS 2008. This control takes rdlc report definitions, client based versions of the rdl report definitions of SQL Server Reporting Services. Designing the reports is largely the same as in SSRS but there are some gnarly issues when using the ReportViewer control which have taken some experimentation and Googling around to overcome. The following relates to ReportViewer in VB win forms.Mostly it applies to asp.net web forms as well.

What is the ReportViewer control?

Well it is just that, a control, and it has to be placed in a form. It is in effect a container for a report. The wizard allows you to pick a report and you can change it later via the properties box for the control but there are issues with data sources after that, which I'll cover later.
If you want to specify parameter values, you have to add those to your form as text boxes or combo lists and pass through the values through to the report. That's where the fun starts.

I've found it's often easier to view the report definition as xml and edit this directly, especially if you need to change the sql definition for the data source, for instance if you've changed the data type of a field.
Filtering

There are two ways of doing this: by passing the filter as a parameter through to the report or by filtering the report dataset loaded with the form.
I never got the first method to work. I might revisit the problem now I have a greater understanding of the ReportViewer control. The control is quite unforgiving to coding errors and I've found that most problems are related to data in one way or another.
In a report which shows overdue actions for a case worker, I wanted to be able to filter the report by Case Worker. The easiest way I found of doing this was to select the case worker from a list on my form and then refill the table adaptor and refresh the report viewer.

Dim awid As Integer = cboCaseWorker.SelectedItem

Me.taCaseworkerActivityReport.Fill(Me.dsReport.spCaseworkerActivityReport, Me.dtStartDate.Value, Me.dtEndDate.Value, awid)

Me.ReportViewer1.RefreshReport()

where, dsReport is my dataset, spCaseworkerActivityReport is a sproc and dtStartDate, dtEndDate are date picker controls on the form, used to specify the period.

Passing parameters

In SSRS, this is pretty easy, you define the parameter in the report, including the fact that its value will be selected from a list and that the data for the list is from a data source, also defined within the report. When you display the report in the browser, SSRS automatically provides the drop down list, the user makes a selection and the selected value is passed to the report.
ReportViewer doesn't do any of this for you. Instead you have to add your own text or combo boxes to the form and pass the selected parameter to ReportViewer with code.
You start by defining the parameters in the report in the usual way. Right click on top left corner of the report and select Report Parameters

In this case I am passing in a single parameter, Period, which is the period for the report, as a text string. This will be displayed in a text box on the report header. Note setting the prompt is largely pointless here - it is never displayed.

In the form's Load event, or a function called by it, add the following code

Dim p As ReportParameter
p = New ReportParameter("", ""))

ReportViewer1.LocalReport.SetParameters(New ReportParameter() {p})
Me.ReportViewer1.RefreshReport()

where in this case is and is my formatted text string (note both arguments are strings, even if the parameter data type isn't)

I never successfully managed to pass through more than one parameter to a report, e.g.

Dim p1 As ReportParameter
Dim p2 As ReportParameter

p1 = New ReportParameter("Param1", "1"))
p2 = New ReportParameter("Param2", "2"))

ReportViewer1.LocalReport.SetParameters(New ReportParameter() {p1,p2})

Sub-reports
Sub-reports are just reports dropped into a sub-report control placed in the parent report. The issue with handling sub-reports in the ReportViewer control is in setting the data source for the sub-report.

If you don't set the data source for the sub-report, when you run the report you will see the message

Error: Subreport could not be shown

Actually, it is necessary to set the data source for each list or table in the sub-report. If the sub-report contained two tables, you would need to set two data sources.
To fix the error message, assume I have a form, frmReport which contains a ReportViewer control, ReportViewer1.

I start by adding an event handler in the Load event of the form

Private Sub frmReport_Load(ByVal sender As System.Object, _
ByVal e As System.EventArgs) Handles MyBase.Load

AddHandler ReportViewer1.LocalReport.SubreportProcessing, _
AddressOf Me.SetSubDataSource

Me.ReportViewer1.RefreshReport()

End Sub

Next define the Sub called by the event handler. In this example, I am assuming the report has two sub-reports, SubReport1 and Subreport2, each with a single table. For each sub-report, I fill a tableadaptor for the sub-report filtering the dataset by the ID of the row in the main report. I then add a new report data source, e.g. "DataSet_vwSubReport1"

The data source name should be copied from the table or list dataset name. Right click on top left corner of table or list and open the properties dialog

Public Sub SetSubDataSource(ByVal sender As Object, _
ByVal e As SubreportProcessingEventArgs)

If e.ReportPath = "SubReport1" Then
    Me.taSubReport1.FillUsingId(Me.dsReport.vwSubReport1, Id)
    e.DataSources.Add(New ReportDataSource("DataSet_vwSubReport1" _
   , Me.dsReport.vwSubReport1))

ElseIf e.ReportPath = "SubReport2" Then
Me.taSubReport2.FillUsingId(Me.dsReport.vwSubReport2, Id)
    e.DataSources.Add(New ReportDataSource("DataSet_vwSubReport2" _
, Me.dsReport.vwSubReport2))
End If
End Sub

This is manipulating the collection of ReportDataSource objects of the report. A ReportDataSource object has a name and a value which is the data source.

Sub-reports - filtering by main report

In the subreport we must a) define a parameter and b) set the table filter to the value passed via the parameter, e.g. for subject activities

a) define a parameter, SubjectId, by right clicking outside the drawing area to display popup

and select Report parameters

b) set the table filter by clicking on the table and right click in top left corner

Select properties to display table properties dialog and select filter tab. set up the filter

The expression to be filtered is the SubjectId field and the filter value is the value of the SubjectId parameter.

Finally, pass the parameter value from the main report to the subreport . In the main report open the parameters property of the subreport control

The parameter name must match that defined in the subreport. In this example, the value is the value of the SubjectId field.

Switching Reports in Report Viewer

The MCPD operates in both English and Dhivehi and reports can be delivered in both languages. I'll blog elsewhere about the issues of dual language, especially when one of the language uses a non-latin character set which runs right to left.
My approach to the dual language problem for reports is to use the same Windows form, with ReportViewer Control for each language setting and to reference different reports for the English and Dhivehi versions.

For each Language the underlying datasource is different, one delivering text values in English and one in Dhivehi.

In the form.designer.vb comment out the lines

'Dim ReportDataSource1 As Microsoft.Reporting.WinForms.ReportDataSource = New Microsoft.Reporting.WinForms.ReportDataSource

and

'ReportDataSource1.Name = "MCPDDataSet_vwRptE_CaseReport"
'ReportDataSource1.Value = Me.vwRptE_CaseReportBindingSource
'Me.ReportViewer1.LocalReport.DataSources.Add(ReportDataSource1)
'Me.ReportViewer1.LocalReport.ReportEmbeddedResource = "MCPD.E_CaseReport.rdlc"

In the form.vb load event or a custom New or elsewhere add lines to clear the existing report data sources

• declare a ReportDataSource variable
• set its name and value
• add this new ReportDataSource to the report
• finally we set ReportViewer Control's rdlc in code as the project name (MCPD) and report object.

Public Class CaseReport

Private mReportDataSource As New Microsoft.Reporting.WinForms.ReportDataSource

'set report and report datasource
Me.ReportViewer1.LocalReport.DataSources.Clear()
If LanguageCode = "EN" Then
mReportDataSource.Name = "MCPDDataSet_vwRptE_CaseReport"
mReportDataSource.Value = Me.vwRptE_CaseReportBindingSource
Me.ReportViewer1.LocalReport.DataSources.Add(mReportDataSource)
Me.ReportViewer1.LocalReport.ReportEmbeddedResource = "MCPD.E_CaseReport.rdlc"
Else
mReportDataSource.Name = "MCPDDataSet_vwRptD_CaseReport"
mReportDataSource.Value = Me.vwRptD_CaseReportBindingSource
Me.ReportViewer1.LocalReport.DataSources.Add(mReportDataSource)
Me.ReportViewer1.LocalReport.ReportEmbeddedResource = "MCPD.D_CaseReport.rdlc"
End If

You must add tableadaptor and binding source to the form.vb{Design]. The easiest way to do this is to select each one in turn using the reportviewer tasks panel

Selecting the report will add the datasource, tableadaptor and bindingsource controls to the designer

Gotchas

Report viewer is quite unforgiving on coding errors and quite unhelpful in indicating the source of the error. In almost all cases I have found the problem to be connected with data. Below I've listed errors I have made which have caused the ReportViewer control to not display the report.

1) If you change the size of a text field in the database, it is necessary to set the property for that field in the dataset designer for the table AND for any view using that field. If a report uses a table or view where the actual field size in the db and the property in the ds designer are different, it can lead to an error the first time the fill adaptor is called (if you step through the code in debug mode, you can see the fill method fail)

2) If the report parameter is set to the wrong type, e.g. if an integer parameter Id, is defined as string (default) instead of integer, the subreport will not display

3) if there is no select permission on the sql datasource (view, table) for the connection string, the sub report will not display. No error will be raised.

Leaving Male' on a High

2010-05-19T19:33:00.001+05:00

After three extremely frustrating days trying to get a connection from the social workers offices to the database, we finally achieved success this afternoon and got 4 PCs set up and some of the users logging in to see their cases migrated from their old system.

We didn't get the chance to set up in the other agencies but that can be done by local support staff next week.

Jim and I fly home tomorrow morning but we're happy bunnies tonight.

The DGFPS Office on Sosun Magu this afternoon

Jim helping one of the social workers find a case on the new system

A message from Laura (who was the unicef project manager for the 2007 scoping exercise) to unicef today to ask if it had crashed yet! Well yes it did but only when Jim logged in, so that doesn't count.

We're Live

2010-05-16T19:35:00.000+05:00

We went live today - well almost. The server is now in a rack at Police HQ and we have the first admin user set up. Tomorrow and Tuesday will be spent touring the various agencies to help set up PCs and get users onto the system. What Jim and I started in October 2007 is finally realised - a national, multi-agency, child protection database for Maldives.

The server about to leave unicef in a taxi (that isn't the taxi!)

The server outside Police headquarters

Oh yes, we did have one small problem. The server rails got missed off the order to Dell!

Translation with an accent

2010-05-15T23:46:00.002+05:00

After a couple of months of working with the translated text for the database (screens, lookup lists, error messages, etc) along with the issue of thaana characters in ASCII and unicode (to be blogged shortly), I've learnt most of the common Thaana characters.

People here have been telling me that dhivehi is written phonetically. Well now I'm going round reading signs in Thaana (to the annoynace of my colleagues) and finding that many of the words are just the English translated into a 'pseudo-dhivehi', sometimes with humorous outcomes.

For instance, the desserts section of one restaurant menu offers arse cream (instead of ice cream) and the Bank of Ceylon's new sign reads Baynk of Ceylon, which has more than a hint of an english colonial past and plummy accents. How nayce.

The bank for toffs

Q: When is A not A? - A: When it's an aabafili.

2010-03-27T14:31:00.174+05:00

I received a spreadsheet from the police with a list of their offence codes in Dhivehi, for loading into the database. All of the lists are structured as a code (hidden from the user) and English and Dhivehi descriptions. The description field displayed in any drop down list depends on which language the application is set to. The user can switch betwen English and Dhivehi, although free text fields must be completed in English. We did at one stage consider allowing either language but it gets impossibly messy when the Faruma font doesn't display latin characters and Dhivehi runs right to left.

Anyway, when I started to look at this spreadsheet in some detail, I noticed a problem. The cell contents displayed as divehi but in the edit bar, i.e. what was actually stored in the cell, they were in western (latin) characters.

The contents of cell A10 are displayed as Dhivehi but input and stored as latin characters.

It's taken a few weeks to fully understand the what, why and how of this, which I think I now do. There are a number of strands to this.

Firstly is the keyboard mapping, that is the correspondance between the characters (and ASCII Codes) physical keyboard and the Dhivehi characters they produce when that language option is selected.

Next, the font we are using, P_Faruma, along with a number of other dhivehi fonts will display the unicode Thaana character set and the ascii character set as the equivalent Thaana characters. A consequence of this is that you can't dispay a latin font based language, such as english with these fonts. Some of the standard Windows TrueType fonts such as Arial, will display only the unicode character set (as far as I can see), so that ASCII 64 is always A, whereas Faruma displays ASCII 64 as the equivalent Thaana character on the dhivehi phonetic keyboard layout (selected through Regional Settings), which is ާ or Aabaafili, the double a sound and unicode character 07A7.

Dhivehi Phonetic (soft) keyboard layout

As I've already noted, this mapping of latin to Thaana characters isn't fixed and depends on the keyboard mapping. My first introduction to this was on a blog by Jawish Hameed, who published the java source for a transcoder he has developed to convert between latin and thaana. When I used his mappings in a VB program to do a similar job, I only got about an 80% match between the latin and Thanna in the spreadsheet data, with quite a few of the filis (vowels) being transposed. My first suspicion was that maybe this was a java thing - perhaps linux keyboards may use a different mapping.  Then again it may have nothing to do with the OS but relate to mapping of the latin characters on the particular physical keyboard to the dhivehi ones. For instance my keyboard is a UK QWERTY layout, which for the main alphabet characters will be the sames as US QWERTY but would be different to say a German or French keyboard. The mapping I describe below isvalid for a US/UK QWERTY keyboard and Windows Dhivehi Phonetic (as far as I know!)

At this point I set about documenting the mapping for Dhivehi phonetic on a qwerty keyboard and once I had modified my VB code for this, I got 100% match between the latin based dhivehi displayed in the spreadsheet cells and the true unicode dhivehi. The mapping is shown below. There are a few special characters to deal with. Dhivehi reads right to left and the Dhivehi equivalents of question mark, comma and semi-colon are mirrored, as are open and closed bracket. In common with arabic, numbers run left to right whilst dates are read right to left and so formatted as yyyy mmm dd.

A big thanks must go to Jaa. The code below is mine not his but the initial mapping in his code set me on the right track.

Some other useful resources are:

http://tlt.its.psu.edu/suggestions/international/bylanguage/thaanachart.html

http://en.wikipedia.org/wiki/T%C4%81na

http://unicode.org/charts/PDF/U0780.pdf

The following is good in VB6, VB.net, Access Basic. This version handles thaana strings without any numbers, whose digits will get reversed. I have another version which does a more complete job, though one thing I have yet to crack is coding for /- used in the representation of currency values.

Function ThaanaAsciiToUnicode(ByVal strIn As String) As String

'copyright (C) 2010 Tony Bennett

'You are free to use the code for personal or commercial purposes as long as you retain the copyright notice

'The pheonetic keyboard layout mappings are:
'h' -> '1920','S' -> '1921', 'n' -> '1922', 'r' -> '1923', 'b' -> '1924', 'L' -> '1925',
'k' -> '1926', 'w' -> '1927','v' -> '1928', 'm' -> '1929',
'f' -> '1930', 'd' -> '1931', 't' -> '1932', 'l' -> '1933',
'g' -> '1934', 'N' -> '1935', 's' -> '1936', 'D' -> '1937', 'z' -> '1938', 'T' -> '1939',
'y' -> '1940', 'p' -> '1941', 'j' -> '1942', 'c' -> '1943',
'X' -> '1944', 'H' -> '1945','K' -> '1946', 'J' -> '1947',
'R' -> '1948', 'C' -> '1949', 'B' -> '1950', 'M' -> '1951',
'Y' -> '1952', 'Z' -> '1953', 'W' -> '1954', 'G' -> '1955',
'Q' -> '1956', 'V' -> '1957','a' -> '1958', 'A' -> '1959',
'i' -> '1960', 'I' -> '1961', 'u' -> '1962', 'U' -> '1963',
'e' -> '1964', 'E' -> '1965', 'o' -> '1966', 'O' -> '1967',
'q' -> '1968',

'plus the special characters
',' -> '1548', ';' -> '1563', '?' -> '1567', ')' -> '0041',
'(' -> '0040', 'Q' -> '65010'

'note that ? , ; ( ) are mirrored in thaana

Dim strOut As String
Dim c As String
Dim i As Integer
Dim j As Integer

'Dhivehi Phonetic <-> QWERTY mappings
Const AsciiChars1 = "hSnrbLkwvmfdtlgNsDzTypjcXHKJRCBMYZWGQVaAiIuUeEoOq"

'The special characters
Const AsciiChars2 = ",;?)(Q"

strOut = ""

If Len(strIn) > 0 Then

'check if any chars in unicode.
'if they are, assume this is unicode dhivehi and skip the conversion
For i = 1 To Len(strIn)
If AscW(Left(Trim(strIn), 1)) > 255 Then
    AsciiToUnicode = strIn
    Exit Function
End If
Next i

For i = Len(strIn) To 1 Step -1

c = Mid$(strIn, i, 1)

'need to do a case senstive instr j = InStr(1, AsciiChars1, c, vbBinaryCompare)

If j > 0 Then
    strOut = strOut + ChrW(1919 + j)
Else
j = InStr(1, AsciiChars2, c, vbBinaryCompare)
    If j > 0 Then
    Select Case j
       Case 1 ',
       strOut = strOut + ChrW(1548)
        Case 2 ';
          strOut = strOut + ChrW(1563)
      Case 3 '?
         strOut = strOut + ChrW(1567)
      Case 4 ')
        strOut = strOut + ChrW(41)
        Case 5 '(
          strOut = strOut + ChrW(40)
       Case 6 'Q
         strOut = strOut + ChrW(65010)
        End Select
     Else
       strOut = strOut + Mid$(strIn, i, 1)
     End If
   End If
Next i
End If

AsciiToUnicode = strOut

End Function

SQL Collation for Dhivehi

2010-03-24T19:40:00.010+05:00

When I originally set up the database back last September, I used the default latin collation for SQL Server, Latin1_General_CI_AI. I knew at the time I would need to consider a different collation for the dhivehi content but then forgot about this issue until the other day when I tried a search against fields holding dhivehi as unicode. My first search returned all records in the table, irrespective of the filter string. For instance searching the dhivehi firstname field for the dhivehi Bart, ބާރޓް
in a table containing records for {Bart, Homer, Lisa, Maggie, Marge} as { ބާރޓް, ހޯމަރ , މެގީ , މާރޖް , ލިސާ } using

SELECT * FROM tblPeople WHERE D_FirstName LIKE N'%ބާރޓް%'

records were returned for all the Simpson family. I was sort of expecting something like this might happen, though if I'd had to put money on it, I would have guessed I'd get no matching records.

By changing the collation on this field to one of the dhivehi set supported by SQL Server 2008, only the record for bart was returned.

The choice of collation took some consideration. The set of divehi collations in SQL Server 2008 numbers over 30 variations:

We can reduce this list by half by using the 100s in preference to the 90s. According to sql 2008 books online:

SQL Server 2008 has introduced new collations that are in full alignment with collations that Windows Server 2008 provides. These 80 new collations are denoted by *_100 version references. They provide users with the most up-to-date and linguistically accurate cultural sorting conventions.

We don't want to use the BIN collations.

Dhivehi doesn't have cases. A blog at http://www.jawish.org/blog/plugin/tag/how+to, referring to MySQL suggests a collation of utf8_unicode_ci, which isn't divehi specific but does indicate that case insensitive is good enough. Also, a blog at http://blogs.msdn.com/michkap/archive/2005/05/11/416293.aspx suggests that case sensitive can lead to problems. So case insensitive seems a good option.

The filis are stored as separate characters rather than accented versions of consonants. The choice between accents sensitive and insensitive probably makes no difference but accent insensitive seems a safe bet.

Kanatype indicates whether the data in the dataset is kanatype sensitive, which distinguishes between the two types of Japanese kana characters: Hiragana and Katakana. This is of no relevance to Dhivehi, so insensitive sounds a good option.

Width isn't an issue but insensitive seems the safest option.

This brings the choice down to

Divehi_100_CI_AI Divehi-100, case-insensitive, accent-insensitive, kanatype-insensitive, width-insensitive

To test this I created the table

CREATE TABLE [dbo].[tblPeople](
[PersonId] [int] IDENTITY(1,1) NOT NULL,
[E_Firstname] [nvarchar](40) NULL,
[E_Lastname] [nvarchar](40) NULL,
[D_FirstName] [nvarchar](40) COLLATE Divehi_100_CI_AINULL ,
[D_LastName] [nvarchar](40) COLLATE Divehi_100_CI_AI NULL)
GO

populated d_firestname with the simpsons o=in dhivehi and repeated the earlier query

SELECT * FROM tblPeople WHERE D_FirstName LIKE N'%ބާރޓް%'

which now just returns Bart's record.

Note, it is important to prefix the search string with N to indicate a unicode string.

A ship the size of an island

2010-03-24T17:20:00.002+05:00

The P &O liner Aurora arrived off Male' last night. At almost 900' long, it's bigger than the next door island of Funadhoo.

The P & O Aurora moored opposite the President's Jetty on Bodu Thakurufaanu Magu, Male'

It's not a pidgin

2010-03-24T14:32:00.003+05:00

There's a road close to the hotel called Filigas Magu (Filigas Street). At the end of it is the only petrol (gas) station on Male'. Initially, I wondered if it was pidgin but it turns out to be just an amusing coincidence. Filigas is a kind of tree. Fili is also the name given to the 'squiggles' above and below the main characters in dhivehi, which represent the vowel sounds.

Not pidgin

Pigeon

Meanwhile back on the database...

2009-12-02T22:06:00.002+05:00

Latest news from the project is that I added another 800 islands to the database yesterday and lost 200 health centres in the process.

The project plan for phase 2 has been redesigned and considerably slimmed down with only two more trips planned and an official launch on May 10'th which is Maldives Childrens' Day.

More meetings with social workers and Dept of National Planning to discuss reporting requirements.

When not in meetings I've spent the last two days finishing the end of project report for phase 1, rather than finishing off the project. I do believe I may have fallen into the unicef way of life.

Coast Guard boat sailing between Male' and Dhoonidhoo

On another note, the restuarant staff at the hotel have exhibited another bout of haplessness at lunchtime when the beef fried noodles I ordered arrived as beef fried rice, apparantly destined for another diner. I wonder if he got my chilli sauce as well since what I got was plain old tomato ketchup!

For no other reason than it's one of four streets on Male' I know the name of

Circular reasoning

2009-12-02T21:46:00.000+05:00

Currently, mosques are the only places of worship in Maldives. I read in a newspaper here from last week that there was a proposal to permit places of worship for other faiths to be built in the country.

However one muslim leader noted that as Islam was the only religion in Maldives it wasn't necessary to allow this, which rather neatly brings closure to the subject.

A mosque on Male'

Abuse of a different kind

2009-12-02T21:36:00.004+05:00

This time, it's abuse of the English language.

I was sitting in the unicef office yesterday and Latheef says to me, "Tony, can you explain what this means"? From the look on his face I could see he felt some disquiet in what he was reading. It was World Aids Day yesterday and he had found the following on an HIV/AIDS website. As soon as I read the second line I could see the cause of his concern.

While male to male sex is often discouraged from a moral or religious point of view, from the public health perspective it is desirable to engage men having sex with men in an attempt to help them reduce HIV transmission; not only between men...

It was the use of the term engage which has been adopted in recent years by many people as a replacement for a much simpler word, talk. The closest definitions of the word engaged for the meaning intended in the about sentance are

To attract and hold the attention of; engross: a hobby that engaged her for hours at a time.

To draw into; involve: engage a shy person in conversation.

But how much clearer it would have been to just write 'talk to'. Somehow by engaging with people it conveys more importance in the engager and what they are saying. I think it is probably meant to suggest that a two way discussion will take place but there is no guarantee that the engagee will actually be engaged.

What of course it does not mean in the context of the above is encourage men to become engaged to other men!

Unicef engaging the wider community

Child Sexual Abuse Act

2009-12-02T20:05:00.002+05:00

There was a new child sexual abuse act passed recently in Maldives.

The act contained many good things to help protect vulnerable children, in particular young girls.

However one article was included which stated that if a girl is married under sharia law all the other articles in the act no longer stood because she is no longer considered a child but an adult.

I think under Sharia law in Maldives the age of consent is12. It might have been increased more recently by legislation in the parliament, but the general feeling here is that the new act may do more harm than good.

So this is one of the reasons we are here and this place is where we doing it from.

Whilst checking on the age for marriage on the web, I stumbled upon this item in another blog

Outraged and Rubbish

2009-11-24T21:56:00.004+05:00

I learnt an interesting fact from one of the unicef guys today. It's 20 years since the introduction of the UN Convention on the Rights of the Child, commonly referred to as the CRC. See http://www.unicef.org/why/why_rights.html.

Since that time every country in the world has signed up to the charter except two. Somalia was one and they signed on 20'th of this month. So the only country which has yet to sign is... The United States of America. For a nation which will go to war in the name of democracy it beggars belief that they are unable to formally acknowledge basic human rights for children.

Rant over*

Otherwise, we have inched forward again today. Still hoping to meet with the head of one of the lead agencies. I'd feel happier if I knew we had some 'senior managemnet buy in' - to use the parlance of the commercial world.

We have a workshop planned for tomorrow where I would like to come away with some agreement on the codes we are going to use in the database and then Jim and I are off on a two hour boat ride to Eydhafushi. The plan was to install a test system with the social workers who I met there last month. We learnt today that they have been posted to an island in another atoll, so we have new people to meet and hopefully they will be enthusiatic enough to want to act as testers and provide feedback. I also want to check if the internet connection speeds are going to be adequate out there, the suggestion being that they may be quite slow.

Jim is working through spreadsheets of codes and there are a few changes to the system going to be needed as a result. We've dropped about 3 pick lists and gained at least one new one. This is living the agile development dream.

The local dust cart, which coincidentally is called Ufili (u fili up, we empty out). Anyway, today, I for one am feeling rubbish.

* In the spirit of citing both sides of the story, the following link offers an explanation of why the the US haven't ratified the charter http://www.unicef.org/crc/index_30229.html

Timing is Everything

2009-11-24T01:20:00.004+05:00

Yesterday's leaden grey skies cleared and woke to more usual tropical paradise weather with clear views from hotel window of dhonis crossing between Male' and airport island.

Fixed more bugs without creating any new ones today. System almost ready to take to Eyadafushi on Wednesday afternoon. We've been given a document by unicef describing child protection processes in use or to be implemented (I'm not quite clear which), which lists categories and sub-categories for abuse, delinquency, domestic violence, behaviour,etc. We have a workshop planned for Wednesday morning with social workers and police to agree on codes to use in the database. These look like the ones to use. Then all we have to do is map the 87 codes from the social workers' existing database and an as yet to be determined number of offence codes from the police database to these new ones.

Malcolm added a logging function to the system so we could accurately measure times for login and to load a case record. The connection from unicef was very slow this afternoon . For a while I thought the server back in Sheffield that we're running against had died. The logging software highlighted what I suspected, which was that the addresses table had got so big with the upload of the social workers' data that weprobably shouldn't treat it as a look-up list any more. We preload the look up data on successful login to the local PC, which saves a lot of data activity everytime we open a screen which references look up data, e.g. in combo boxes. However, the addresses table now has around 6000 records in it and is likely to double in size when we pull in data from the police database. We still might preload these data but do so as an asynchronous background activity while the user is viewing their current tasks and reading any new messages (or making a coffee!)

Last night's 7 hours sleep doesn't look like getting repeated tonight. The aircon in the room is noisily struggling and failing to keep the temperature down.

Morning rush hour traffic on the way to the airport and Hulhulmale

Match Result

2009-11-23T01:46:00.003+05:00

Maldives lost but locals are stoic in defeat. It was only a friendly before the South Asian Football Federation Cup, which Maldives won last year, beating India 1 - nil.

Huge rain this morning leaving streets in central Male' flooded under a few inches of water, which had drained away by the afternoon.

Almost got to meet the head of one of the lead agencies today but she somehow managed to get into another meeting in the hour which elapsed between her agreeing to us and our arrival!

Otherwise a somewhat frustrating day with .net, fixing some bugs and creating a few new ones. Also some distrurbing news that the internet connection from the islands is only likely to be 128kb/s and in any case the maximum available (at extra cost) is only 512 kb. Running the system against a database on our server back in UK from both the hotel and unicef office has proved disappointingly slow to load and save a case record and its associated data. So much for vb.net architecture using disconnected datasets. I'm starting to wonder just what sql and data are being moved about behind the scenes.

A disconnected dataset

Mango Milkshake Island

2009-11-21T23:55:00.030+05:00

Finally got to sleep around 3:30 this morning. The valerian must have kicked in because I was woken at 7:30 by the alarm and feeling drugged. Decided to give going to the gym a miss and have an extra half an hour. Next thing it was 8:25 and I was due to go to breakfast with Jim at 8:30. Fumbled some clothes on just as he knocked on the door and wobbled, dazed, downstairs. I'm going to take one earlier tonight.

There is a lot of excitement on Male' today. We have a big football match, a friendly between Maldives and Hyundai, Korea, and the Korean team are staying in the hotel. I think they may have Germanic origins as two of them were washing the team kit in the shower by the pool this morning and then left it out to dry on the sunbeds.

Jim and I spent a couple of hours working and then had a swim and a short siesta before heading for the ferry to 'mango milkshake island', more commonly known as Vilingile. The ferry terminal is at the opposite corner of the island to the hotel and I was able to show off my newly found navigational skills to Jim who was suitable underwhelmed. The ferry is 3 ruffiyaa each which is about 6p so I gave the cashier 2 5Rf notes and she gave me 3 tickets and 1Rf change but that was sorted amicably. I'm thinking it's time I learnt to count from 1 to 10 in dhivehi. New task for this visit.

Dhonis in Vilingile harbour

Container ship waiting to unload

Beach on Vilingile

Lunched in resataurant by ferry terminal, surrounded by plants and next to a small fish pond with 1 fat fish, 1 long thin fish and a terapin. Served by an Indian waiter who wanted to come to London, which I said was ok because I lived in Sheffield. Lunch was beef fried noodles with lashings of chilli sauce and a Mango Milkshake (of course), which was as good as the previous one I'm pleased to report.

Then we walked along the dock and I photographed boats and we sat on wall and watched the world go by, when a couple of police who looked about 18 came up on a scooter for a chat. They wanted to know why we were here and we told them and asked about the big match and I asked who was going to win and of course they said Maldives, and they scooted off smilins. We saw them again later and they gave us a wave. It's always good to keep in with the local law enforcement officers.

Mosque on Male' near Vilingile ferry terminal

Boduthakarufaanu Magu (Marine Drive), Male' Note the Tetrapod sea defences.

New building on Male'

Then we got the ferry back to Male' and completed the circuit of the island (just about) by coming past the docks and markets. Distinct absence of rambutans (a fruit similat to lychees) - most disappointing. Enough scotch bonnet chillies to set the world on fire. It's just the same as in West Africa. Why does anyone need that many scotch bonnet chillies?

Apparantly the Koreans are playing tonight and tomorrow, so the dining room will be full again in the morning.

Return to Male'

2009-11-20T23:50:00.000+05:00

Arrived on Male' with Jim this afternoon. The plane was pretty empty on the Dubai to Male' leg and I took the opportunity to stretch out on the 4 seats between the aisles for an hour, which was wonderful. Other than that the journey was notable only for the tedium.

Checked in at Mookai again, unpacked and went for swim on the roof and listened to call to prayer from mosques across the city. In two of them the muezzin seem to be harmonizing. Took photos of sunset. Just georgeous scarletts and crimsons and yellows.

There was a slight breeze this evening and so there is none of that sultry, close feel characteristic of the tropics. The hotel restaurant finally managed the chicken curry, which was worth waiting for and I joked with the waiter about whether the chef had been paid yet. Maybe he has and so the full menu is back. There appears to be an overseas football team staying at the hotel.

Time to sort out sleep patterns. Bought some herbal sleeping pills in Manchester whose active ingredient is mainly velerian. I'm sure that the stuff used in gothic novels to knock out the clueless visitor to the remote house on a dark and stormy night! We'll see if it works. Day off tomorrow anyway.

Bothersome Datasets

2009-11-16T13:06:00.006+05:00

Of the half dozen agencies who will use the CPD, there are two who will provide the bulk of the information to be stored in it. Both already have databases which record information relating to child protection, delinquency and other behavioural issues and gender based violence. One of the agencies plans to replace their existing database with this new one, whereas for the other, the CP database will run in parallel with their own, holding more detailed information about Child Protection cases.

Both database have been in existence since the early 1990s and the suggestion is that about 75% of the cases in the one database are also recorded in the other. The remaining 25% of the former would be cases of say domestic violence or bad behaviour which were only investigated by one agency. It seems possible, though it has yet to be confirmed, that there may be cases in the other agency's system which should be migrated to the CP database and which have no equivalent record on the other system.

Merging datasets is a challenge at the best of times; where fields overlap you have to make a decision which data to use, based on factors such as quality or completeness. The task which lies ahead would be daunting enough if that were the only issue. We have two additional challenges to overcome:

1) One dataset is recorded in Dhivehi, whilst the other is in English

2) There is no common reference. Each database has its own case reference number for uniquely identifying a case, so the same case in the two databases have different reference numbers and neither holds a cross reference to the other. The people in the case, or at least some of them will be the same. However, while one system has links to the National Registration database and records the National Id of case subjects, the other system doesn't, so we can't use that to match people across the two datasets. We could attempt a match on names, although variations in spelling are not uncommon, and one would expect there to be a start date for the case which would at least be similar. Nevertheless, it looks like this is going to require a large amount of matching cases by eye unless there are documents or some other information providing a common unique identifier.

Currently I think the best option is to import the two datasets independantly and stored them with their legacy case identifiers, without making any attempt to match cases. It's probably more imprtant we can match people at some stage as this will be allow users to see what cases a person has previously been a subject of.

The 2007 workshop on Bandos

Call to Prayer

2009-11-02T02:01:00.007+05:00

I love the Islamic call to prayer. It occurs five times a day and I find the chanting of the muezzin immensely relaxing. The first time I heard it, I was workling in Saudi for a few days, installing a Computer Aided Design system. Things had not been going well. We couldn't get the graphics on the screen properly (this was in the mid-80's when this stuff was cutting edge and difficult). We were trying to debug code when in fact it was a hardware setting on the graphics card. And then the call to prayer started and for a few minutes everytning in the world seem tranqil and at peace.

Whenever I hear a call to prayer now, I think back to that moment. These days, many mosques seem to use a recording, which isn't the same as hearing the voice of a real person singing live.*

There is a mosque close by the hotel and this video clip was taken from the hotel roof last Friday, when we were once again grappling with some software issues, whilst sitting by the pool. Different place in the world, different programing language, different appliction, same frustrations. Plus ca change....

*Strangely enough, just as I was finishing this post, there was an item on Radio 4 breakfast news about the use of recordings of the call to prayer from mosques in Cairo.

Eydhafushi, Baa Atoll

2009-11-01T22:15:00.004+05:00

Up early to get launch from jetty #1 outside police HQ to Eydhafushi on Baa Atoll, about 100km to the NW of Male'.

Jetty No. 1, looking towards the main docks and markets

Leaving Jetty No1, Police HQ and the Presidents Jetty in the background.

The sea was pretty calm all the way, even when we left the protection of N Male' Atoll. I feel slightly uncomfortable in the main channels between the atolls. The ocean drops away to around 2000' in depth and it feels like a very small boat to be in such a large expanse of water.

Our high speed launch

Went past Bandos. Saw a dolphin. Thought about a few problems relating to the database. As we approached Eydhafushi we saw ominous dark clouds ahead. As we entered the shallower waters off the island, there was quite a bit of chop and it felt like the boat was being driven over large cobbles.

Dhonis, traditional Maldivian boats

Mannan was waiting for us when we arrived and after dropping bags of at guesthouse, we walk round to police station for a second breakfast: roshi, curry, tuna/whitefish and of course betelnut to finish.

Discussed training and roll out to police and social workers and agreed we needed to rework the whole project plan. We need to train the two groups together to give them the opportunity to develop new processes. We can't train until we have migrated at least the current case data from social workers system and so it looks more like the training will now be put back until early 2010.

Met with two of the social workers from the island and showed them the system so far, to what one can interpret as great enthusiasm (emotion can be hard to detect in Maldivians)

Checked out training facility at Teachers Resource Centre and learnt how to write Baa Atoll in Dhivehi.

Baa Atoll written in Dhivehi. The text runs right to left with vowels (filis) above and below the consonants. Baa is also one of the letters of the Thaana character set.

We had been invited round to the police station in the evening for a barbeque. Going past the harbour area on the way there, we watched crabs scurrying about and lightening across the water.

Eydhafushi harbour at night

It was halloween back home and on Eydhafushi we watched the silhouettes of fruit bats fly across the face of an almost full moon.

Ryan and Nancy sitting in traditional Maldivian outdoor seats

During the night there was a typical tropical rainstorm where the gods just emtptied bath loads of water onto the ground. Sunday morning was clearer skies, fierce heat and a lot of humidity.

More dhonies in unfeasibly blue water

A brief visit to see the social workers at 8am, then breakfast at the police station and the boat back to Male' at 9 o'clock.

Eydhafushi main street

We passed a pod of dolphins shortly after leaving Eydhafushi and as we got closer to Male' we made a small detour to avoid a storm, when things started to get quite exciting.

Grand Friday Mosque, Male'

A searching problem

2009-11-01T20:27:00.002+05:00

In any database system which records details about people it's important to be able to search against a person's name as the primamry means of pulling back their record. A large database may contain records for people having the same name in which case you would then refine the search with other details such as date of birth.
In English and most western languages there are few varitions to the spelling of a name

In the Dhivehi language there can be subtle variations to the spelling of a name, which can make searching for an exact match a problem. To begin with words in Dhivehi can be written across five lines, with the consonants in the middle, vowels above and modifiers or accents above or below. A particulr problem which was described to me this week by one of the police IT guys is the use of the dot. The dot was introduced into written dhivehi about 30 years ago and is used to subtley modofy the sound of a letter, think of the grave and acute accents in French but more so.

If they apprehend someone and take down his name incorrectly by missing a dot, they may not find him in the database. For instance if they apprehend John Smi.th but search for the equally common name smith, either he will not appear in the results list or they will match on the wrong person. John Smith may have no previous convictions, whereas John Smi.th may be a serial offender. At the very least the police end up creating two records for the same person and in the worst case, he gets a sentance less than he otherwise might.

This problem isn't unfamilar in databases of English names and a way round it is to do wildcard searches where you include a character which matches against any single character or string of characters. So a seach for say Sm*th or even Sm* would match on Smith and Smi.th. Clearly the less specific the search, the greater number of false matches. if we just searched for * we would find everyine in the database!

The problem is further compounded in Dhivehi when the first and last names may be transposed, although once again this isn't entirely unfamiliar in UK databases which store the names of, in particular, people of asian descent. It is normal to store first, middle and last names in separate fields in a database, in which case searches have to match on say

Firstname = 'Smith' OR lastname = 'Smith'

If we have to include middle name(s) in the search, then the search becomes correspondingly more complex and will take longer.

Pigeon and Pig.eon

For this reason, the police system stores the full name in a single field and they have suggested we do the same, which is an idea that I'm not enirely comfortable with. In a large database, doing wildcard searches instead of exact match searches is going to impose a performance overhead. Then again we are only looking at a dataset of around 1000 cases a year since the early 1990s.

Every Maldivian has a national ID number, assigned at birth and carries an ID card which inlcudes their full name in Dhivehi and English, their 'permanent address', which is the address when they were firsr registered, a photo and more recenlty, fingerprint data. The national ID database stores names as first, middle, last and common name, the latter being the name the individual commonly uses. So the police approach of a single name field is also at odds with the national ID database.

The final consideration in determining a solution to this problem relates to the merging of data from the police and social worker databases, which I intend to blog about separately.

It's the weekend

2009-10-30T23:44:00.001+05:00

The weekend here starts on Thursday night, so Friday we spent the morning by the roof top pool, fixing bugs in the system (oh so many still and missing functionality). Ryan and I were getting a bit fraught with various annoyances in .net when the call to prayer started up from mosques (miskiy) across the island.

View from the hotel roof (by the pool)

After a swim to unwind we set off to get the ferry to Vilingile a couple of miles across the water. The ferry teminal is at the opposite corner of the island from the hotel but I'm starting to get a feel for the place now. - when Jim and I came out two years ago, we didn't venture out that much. We had planned to eat at Salsa 205, a place on Ameenee Magu a few minutes walk from the UN building but it was shut when we got there, so we got lost round some back streets and found a nice little resaturant for more stir fired noodles (oh the pounds are piling on)

The ferry teminal was easy enough to find. 3 Ruffiya each (22 to the £) and there was a boat waiting.

Dhoni or water bus

Inside the Vilingile Ferry

It was stupidly hot this afternoon and with almost no air movement, we ambled round the island looking at views of Male', small, almost translucent crabs scurry at breakneck speed across the sand, stripey fish and blue fish with yellow fins and a long pipe (?) fish plus lots of non-descript tiddlers.

Me - taking the afternoon off!

Took a lot of photos of boats (they're such interesting shapes and colours)

Fishing Dhonies

The new company boat :¬)

and stopped at the cafe next to the ferry terminal on the way back for a yummy mango milkshake.

The Mango Milkshake

We arrived back on Male' at dusk and walked round the 'coast road' - the one with the long unpronouncable name past the docks and markets (in any other country it would be scary place at night but it doesn't feel that way here.

Male' from Vilingile

We are supposed to be getting a motor launch to Eydhafushi on Baa Atoll tomorrow morning, 100km and 2hrs away. However, we got a text this evening to say the weather wasn't good. If it's too rough we'll have to cancel, which will be a pity because we were going to meet someone we worked with closely two years ago.

Days 1 and 2: The Journey to Male'

2009-10-27T01:36:00.006+05:00

The trip did not start well. I finished packing, which took the form of arranging the contents of various piles of stuff on the living room floor into the spaces afforded by a largish suitacse and small rucsack. I left the house at 2pm, giving plenty of time to get to the station for the 15:10 train to Manchester airport. I waited at the bus stop for 20 minutes, thinking it slightly odd that no bus had gone past, when a family walking by informed me there was a bus strike. Brief panic set in and I rushed back to the house, could get hold of a taxi and decided to drive into town and leave the car at the office. I got about a quarter of a mile and thought how much better ot would be if I dropped the spare keys off with a friend so she could collect it the next day and park it outside her house. So I turned round. Finally setting off again I got to the station with time to spare to meet the other two. The train journey, check in and take-off were uneventful. After an evening meal and watching the remake of Pelham 123, I set about catching up on some sleep.

I woke to find a clear night sky and the lights of villages below, including some very large bright red ones. The in-flight information screen showed we were crossing Iraq, to the east of Baghdad. As we flew more or less directly over one of the red lights I could see it flickering and realised they were flames, the burn off of unwanted natural gas from the oil wells. Also, alternately red and white sheets of lightening travelled between and along cloud layers in the far distance, over towards the Zagros Mountains of Iran. At least I hoped it was lightening and not a resumption of large scale military activity.
As we flew south, the sky above us cleared more and I could see stars and one especially bright one, almost certainly not a star but Venus, appearing balanced just above the port wing. For a while I had thought it was a light on the wing until we banked slightly and it momentarilly parted company with the leading edge. The colour of the sky was starting to change with the glow of sunrise and from the east round to the south were bands of light going from black at ground level to crimson to orange, light blue, dark blue and finally back to the black of the night sky. The aircraft wing was silhoutteted running parallel to these bands with Venus poised just above it and other points of light, possibly the big dipper but standing on it's handle, punctering the jet black of the remaining night. There was a sudden burst of noise from the aircraft engines and we turned more or less due south to fly parallel to the River Tigres, and passed over the lights of Faw, where this great and historical river flows into the Arabian Gulf.

Dubai Airport Passenger Terminal

We landed at Dubai, bouncing somewhat erratically along the runway as the pilot over-corrected and re-corrected for the cross wind. I've always felt that the only dangerous bits of flying were at take-off and landing, notwithstanding a bird strike or other mid-air collison, but on balance I dislike the process of reconnecting with the ground the most.

Dubai airport is stunning in size and architecture, the main passenger terminal comprising a massive oval tube, presumably symbolic of an aircraft wing, stretching for a good quarter of a mile, if not twice that distance. Inside, maps showing the layout of the shopping and eating areas show indicate the time in minutes it will take from the observers current position to various places: 2, 5, 10. Probably 20 to 25 minutes end-to-end.

Duty free coffee pots. Well I do drink a lot of coffee!

We stopped at a Costa Coffee for a drink and snack and then perused the duty free for an hour, deciding that the prices for most of the electronic goods, cds , dvds etc, were the same as back home and buying nothing, which is my idea of a succesful outcome to that sort of shopping trip.

The take-off for the second leg of the flight to Male' was uneventful. We dozed, we ate, we dozed some more. The landing at Hulhule (airport island) was even more erratic, not helped by watching the view from the forward camera on the TV screen where our high speed deviations from the line running down the centre of the runway were only too apparant. Getting though Customs took over an hour, with an additional form to complete relating to the risk of swine flu, which on Male', one of the most densley populated regions of the plannet having 170,000 people living together in and area of two square kilometers, would be hard to control, if not catastophic.

Male' Airport

We emerged from the airport to see no sign of anyone to meet us and after failing to make contact with unicef by phone or to get a wireless internet connection to email from the laptop, I decided we should hop on a Dhoni for the five minute crossing to Male'. The journey cost 1US$ per person and we landed at Pier 10, opposite the old Nashandura Palace hotel where Jim and I had stayed 2 years ago and which was now closed and empty. Our hotel was two minutes walk from the jetty and check in was easy enough, although the rooms were a little disappointing compared with the other place. Tired and sticky from the travel and struggling to get the air-con to work, I wondered how I was going to spend around 18 weeks here in the next 8 months, some of it alone, and I wondered if I had been over-enthusiastic about the project. I eventually got an internet connection sorted out so I could email unicef and had started to unpack when the phone rang. It was Naeem from unicef. He was under the impression we were ok with making our own way from the airport and had phoned the hotel probably minutes before we arrived. It was a pity they didn't think to tell us.

We showered and met up at 7pm for dinner, which proved a challenge to stay awake over. By eight I was in bed and dozing but this was intermittent as there was a traditional Boda Beru (literally big drum) band playing across the street. Now those of you who know me will know that I am quite passionate about hand drumming, so the steady pounding bass lines, the kunk-e-tunk so familar across West Africa, where it's known as DunDunBa, were not unpleasant. Indeed on any other night I would have been across the road in the audiance but on this particular night, after probably less than eight hours sleep in the last forty-eight, it was not so welcome.

Eventually, less tired, I decided to open the window to hear the music more clearly and that was when I felt for the first time that day that I was glad to be back. The smell of the warm, moist night sea air, the beacon across the water on Hulhulmale, airport island, the sounds of the dhonis moving in and out of pier 10, the lights on the water, the 100cc motorbikes racing round the ring road of the long and unpronouncible name and the cooler night breeze. Yes, perhaps this wouldn't be so bad after all.

View from the room at the Mookai Hotel, Male'

And then a western style guitar band struck up and the mood was lost, so I did the the only English thing possible and went to boil the kettle for a cup of tea. Only there were no tea bags in the room's brew kit, only coffee. Oh well, there's always some chilled water in the minibar, except that the minibar wasn't switched on and there were was no water in it, only bottles of warm coke and sprite. I gave up!

Postscript: the band continued accompanied by the screams of girls until 1am, when finally silence fell on Male'.

Background

2009-10-24T19:30:00.005+05:00

Two years ago I was commissioned by The Maldives Country Office of Unicef to undertake a scoping exercise and write a specification for a multi-agency, national child protection database for the Republic of Maldives. Over a two month period Jim and I met with workers involved with child protection cases from many government agencies and in particular the Family Child Protection Unit of the Maldives Police Service. We were whizzed around in high speed police launches and flown to islands in more distant atolls in the Maldives.

Two years on and we are back, this time to build and implement the database system and train its users who will working from islands spanning 500 miles across the Indian Ocean.

This is emphatically not a holiday (!) although I consider it to be the experience of a lifetime and one of the more important and rewarding projects I have worked on in my professional career.

Male', capital of the Maldives. 2 square kilometers in area and home to 100,000 Maldivians and 70,000 migrant workers, it is one of the most densely populated places on the planet and under the greatest threat from rise in sea levels due to global warming.

As a recent demonstration of the seriousness of the situation, the Maldives government held an underwater cabinet meeting.