Write a routine, in your favorite language or even a mock language, without using an intrinsic function to solve the problem for you, to reverse the contents of a string. The routine should take an input of a string and output a string. For example, if the input is “hello” the output should be “olleh”.

I inform the candidate that it is a simple problem and invite them to talk through their logic as they write their solution on the whiteboard. I let them know it does not have to be syntactically correct; I am more interested in their approach.

It is enlightening, to say the least, to watch interviewees attack this problem. I’ve actually had so many people fail that I’ve been questioned if my “test” was too difficult. I understand there are abnormal pressures in play due to the setting but employees should ultimately perform decently under some pressure.

Silently I watch and listen to the candidate attempt to solve the problem. When the candidate is finished I usually ask, with good reason, are you certain this solution is correct? The response is often a “yes.” This usually gives us ample opportunity to talk about different strategies and try to understand why it was coded in a particular method.

I notice lots of issues that should not occur; for example, the candidate:

• does not understand the chosen language or it’s string libraries.
• has issues with zero-based arrays depending on chosen language.
• overly complicates the approach and never sits back to re-evaluate.
• can not come up with a solution.
• chooses a complicated approach involving math and gets the math wrong.
• uses too many variables.
• writes unnecessary code that essentially does nothing.
• can not validate the output of the routine they wrote correctly.
• writes more than ten lines of code to solve the problem.
• does not know how to swap values in to variables properly.
• … and many more.

I’d appreciate others offering this simple test and sharing your results with me and everyone else. I’d also like to know your thoughts the validity of this test and how you would let the results affect your view of the candidate.

Before reading on to see my solution, try it for yourself. Let me know how you do.

Here is my solution:
static string reverse(string s)
{
string r = "";

return r;
}

Can it be more simple? Let me know?

Transaction Check Error:
file /usr/lib/libsensors.so.3 from install of lm_sensors-2.10.6-1.fc8.i386 conflicts with file from package libsensors3-2.10.5-52.fc6.i386

When I try to remove lm_sensors, a bunch of dependant packages want to be removed, too (but that is not what I want).

I noticed there was a nagios plugin that was being updated called nagios-plugins-sensors. Removing that allowed me to successfully perform a yum update!

So, I typed yum erase nagios-plugins-sensors and yum removed the following:

=============================================================================
Package Arch Version Repository Size
=============================================================================
Removing:
nagios-plugins-sensors i386 1.4.8-9.fc8 installed 1.1 k
Removing for dependencies:
nagios-plugins-all i386 1.4.8-9.fc8 installed 0.0

Success!

This version eliminates the need for the expensive MC-35 and even the special 10-pin connector. I’ve opted for a quick GPS receiver and encased everything in a small black box with a flash shoe mount. Because it is powered through the camera’s power source, a switch on the side of the box turns off the GPS. Version 3 should include a battery, rechargeable through a USB connection.

Read on for directions on how to create your own GPS.

Parts List

1. 4.7Kohm resistor (x 2)
2. 10Kohm resistor
3. 2.2Kohm resistor
4. NPN Transistor
5. IC PC Board
6. SPST Micromini Toggle Switch
7. 3×2x1″ Project Enclosure Box
8. USGloblSat EM-406A GPS
9. Remote Cord for Nikon D200 - find a cheap (<$10) one on eBay
10. Small cable ties
11. Velcro with a sticky mount
12. Epoxy stick

Tools Needed

1. Soldering iron & solder
2. Wire cutters
3. Needlenose pliers
4. Exacto knife
5. Patience

Overview
It’s quite a simple project; just four resistors, a capacitor, the GPS, a switch and some soldering. The goal is to make it tight so it fits in a small enclosure that can be mounted on the camera. I prefer to do a dry run without any soldering to see how it all fits together; I would suggest you do the same.

Disclaimer
Although I have used this successfully with my camera, I cannot guarantee that it will not harm your camera, even if all directions are followed exactly. By constructing this GPS following these directions (or even using them as a basis for your own project), you do so at your own risk and agree to hold me harmless from any damage done to your camera.

Wiring Diagram
The wiring diagram depicted above is the top view of the PC board. My handwriting shows four resistors (r1, r2, r3, r4), a transistor (with emitter, base, and common leads), Red, Black, and White wires, and the six leads to the GPS (number 1 through 6).

Step 1
Attach the four resistors to the PC Board as shown.

1. R1 - 4.7K ohm
2. R2 - 10K ohm
3. R3 - 2.2K ohm
4. R4 - 4.7K ohm

Step 2
Attach the transistor to the PC board, carefully placing the base, common and emitter leads in their proper holes on the board.

Step 3
Attach a wire (preferably red) between the PC board and the switch. This will be used to cut off power to the GPS.

Step 4
Cut the switch off of the cheap remote cord to expose the measly three wires in it. If you happen to have a 10 pin connector with ten wires then you will have to do things a bit differently.

Attach the red wire to the other terminal of the switch. Note the picture depicts a SPDT switch and has an additional terminal. You can choose to leave the switch out as it is a bit obtrusive on the box but doing so will keep the camera active and drain the battery quickly. Alternative, you may prefer a switch with a lower profile such as a slide switch.

Attach the black and white wires to the PC board. The location of the white is depicted with a “w” in the wiring diagram although the picture should help to clear up where it goes.

Step 5
Take the six pin cord for the GPS and cut off one end. I made the mistake of cutting it in half, giving me a much shorter span of wire to work with.

Attach the black wires, numbered 1 through 6 to the locations on the PC board. Wire 3 is connected to the emitter and wire 6 to common of the transistor. The other four wires should be readable.

Step 5a
Solder the connections on the underside of the PC board, if you haven’t done so already. I like to fit everything together first and when I’m happy, do the soldering and then snip the extra wire. You can see from the image that I am not an expert solderer; just make sure that the solder does not stray.

Step 6
Attach the GPS to the PC board using velcro, connecting the white 6-pin connector to it.
Wrap a wire tie around the wire and through a hole in the PC board, make a tight connection to provide strain relief. This is a very important step because if it is not done, one tug on that wire will disconnect something on the PC board making the GPS inoperative.

The device is finally beginning to take a recognizable shape!

Step 7
Create two holes on the long side of the enclose, one to fit the wire through and the other to hold the switch. Make the hole for the switch the size specified for the switch. Remember both the switch and wire will be on the same side as the 10-pin connector on the Nikon body with the switch to the front of the box. If the holes will be different sizes, pay careful attention to where the holes are made.

Step 8
Carefully insert the PC board, wire, and switch into the enclosure. It should fit in there tight. Make sure to secure it to the box if there is anything loose. You can screw the top cover onto the enclosure; we are almost done!

Step 9
This is the tricky part: making the connector. With an exacto knife, strip off the rubber on the 10-pin connector of the cheap remote cord. The goal is to get the plastic piece with the 10 holes and three pins. With patience, the rubber can all be removed and the pins extracted. The pins must be located in different holes than their original location. You’ll want to cut the wire to a size that would ultimately fit between the enclosure and the camera with just a little bit of slack.

Ultimately the wires are connected to the pins in the following sequences. This image is looking into the barrel of the connector.

1. White wire - pin 1 (receive)
2. Red wire - pin 3 (power)
3. Black wire - pin 7 (power ground)

You’ll want to hook GPS power to pin 3 on the camera which always has +5V while the camera is active; note that by keeping the GPS on, this pin will continue to have a +5V supply as receiving the NMEA data keeps the camera active. Using pin 2 to power would keep the power to the GPS flowing even while the camera is off and this would be sure to drain your battery quickly.

It takes a bit of patience and luck to extract the pins, clean them, reinsert in the correct location and solder the wires to the pins. Be careful as the pins are often numbered looking into the connector although you will be soldering on the outside of the connector.

Step 10
Once you have confirmed everything is correct, plug it into your camera and turn the camera on, making sure the switch on the box is in the on position. You should start to see the flashing GPS symbol on the display of the camera within a few seconds. If you don’t, turn off the camera and disconnect the connector and recheck your work. After a bit, the GPS symbol will stop flashing and will be solid (you may need to have a clear view of the sky for a lock to be made). At this point all we need are a few finishing touches.

Carefully disconnect the module from the camera. To make the connector secure, fashion some epoxy in the shape of an elbow connector around the bare wires and the 10-pin connector making sure to secure the pins in the connector. Make certain to leave enough room on the connector to plug it into the camera (about 1/4″). Also, make the elbow bend such that it lines up with how you want your cable to flow. Hopefully you can do a prettier job than I did!

Finally, to make the mount, fashion one out of a hot shoe adapter or take the mount from the bottom of an old, broken flash and mount it to the base of the case. This will give the GPS a place to sit while taking photos. You may need to get creative with how to attach it. The next version will need to add a flash mount on top of the GPS so we can use the GPS and flash at the same time.

Good luck! I look forward to hearing your results and would love to see pictures of your projects.

The failing portion was galleryadd.pl and it was giving me:

Error: incorrect username/password

The resolution is simple: plog-remote.php makes a reference to a $config global that has not been defined. By including plog-load_config.php the gallery remote functionality works!

To fix, add:
require_once("plog-load_config.php");

after the require statements in plog-remote.php.

Being a new father to twins and often pondering the potential cost of education has guided my thoughts on this podcast. We desperately need a shift in the education paradigm; I’m not sure I’ll be able to offer a decent education for my children considering the current college trends. I imagine removing the walls of the learning institutions and replacing them with social networks. And as technology removes the language and distance barriers, social networking can be the driving force behind sharing knowledge - there are so many bright people all over the world.

I also thought about the Borg (Star Trek’s Next Generation version of a highly functional social network) and how a primary goal is to get smarter though collective thinking. Social networking can be used as the medium to add brainpower to solve problems. I would hope in the years to come that the focus of social networking is to share information allowing us all to get smarter.

I wasn’t able to find any utility to move the tasks to trac so I decided to write my own, mostly as a collection of SQL code and offer them for those in a similar predicament.
I am using a trac installation with SQLite and activeCollab running from MySQL. I also have added two custom fields: Vertical and Application. Ad I’m not bringing any attached files.

Also worthwhile to note is I used mysqldump to move the data from MySQL to SQL Server so I could do development there. Some changes to the code below may be necessary to get this to run on MySQL.

The migration consists of generating a bunch of insert statements to populate the SQLite database.

There are a few support routines I created to help the migration.

SQLite3_DateToJulian converts a datetime value into a Julian int suitable for a SQLite int column.

create function dbo.SQLite3_DateToJulian (
  @Date datetime -- Date to convert to Julian
) returns int -- Julian date. # days since 1900-01-01
as begin
  declare @val int
  if (@Date is null)
    set @val = 0
  else
    set @val = datediff(second, '1970-01-01', @Date)
  return @val
end

The next function converts an activeCollab user id into a username for trac. We are using SSPI to map our network logins to a trac user automatically. Our user names are the first character of our first name followed by the last name. Yoy may have to update this as necessary.

create function dbo.acuser (@id int)
returns varchar(50)
as
begin
  declare @val varchar(50)
  select @val = upper(left(username, 2)) + lower(right(username, len(username)-2)) from ac_users where id = @id
 
  return @val
end

The last function formats a string value into a string literal.

create function dbo.strout (@s varchar(max))
returns varchar(max)
as
begin
  declare @res varchar(max)
  if (@s is null) set @s = ''
  set @s = replace(@s, '''', '''''')
  set @s = replace(@s, '’', '''''')
  set @s = replace(@s, '\r', ' ') -- may want to convert this CR/NL translation to something else
  set @s = replace(@s, '\n', '')
  return '''' + @s + ''''
end

There are four trac tables I populated to hold the tickets, comments and custom fields. I always start from scratch so it is necessary to delete the records in each table first.

select 'delete from milestone;'
select 'delete from ticket;'
select 'delete from ticket_custom;'
select 'delete from ticket_change;'


I translate the activeCollab milestones to trac milestones but use the trac milestone name as the Project Name followed by the Milestone name (as my milestones were not unique across projects).

select 'insert into milestone values(''' + p.name + ' ' + m.name + ''', ' + convert(varchar, dbo.SQLite3_DateToJulian(m.due_date)) + ', ' + convert(varchar, dbo.SQLite3_DateToJulian(m.completed_on)) + ', ''' + convert(varchar(max), m.description) + ''');'
from ac_project_milestones m
join ac_projects p on p.id = m.project_id


Tickets are tasks attached to task lists. For us, the task list describes a sub component and we pump this into the keywords field. We could add a third UDF but decided against it.

select
  'insert into ticket values (' +
  convert(varchar, t.id) + ', ' +
  dbo.strout('defect') + ', ' + -- defects by default, could add more logic
  convert(varchar, dbo.SQLite3_DateToJulian(t.created_on)) + ', ' +
  convert(varchar, dbo.SQLite3_DateToJulian(t.updated_on)) + ', ' +
  dbo.strout('Templates') + ', ' + -- Component can (and should) be genericized
  dbo.strout('') + ', ' +
  dbo.strout('major') + ', ' + -- major by default
  dbo.strout(dbo.acuser(t.assigned_to_user_id)) + ', ' +
  dbo.strout(dbo.acuser(t.created_by_id)) + ', ' +
  dbo.strout('') + ', ' +
  dbo.strout('0.9') + ', ' + -- the version
  '''' + isnull(p.name, '') + isnull(' ' + m.name, '') + ''', ' + -- milestone names are project names followed by milestones
  dbo.strout( -- set the status
    case
      when t.completed_on is not null then 'closed'
      when t.assigned_to_user_id > 0 then 'assigned'
      else 'new'
    end) + ', ' +
  dbo.strout(case when t.completed_on is not null then 'fixed' else '' end) +', ' +
  dbo.strout(t.text) + ', ' + -- summary and description are one in the same
  dbo.strout(t.text) + ', ' +
  dbo.strout(l.name) + ');'
from ac_project_tasks t
join ac_project_task_lists l on l.id = t.task_list_id
left outer join ac_project_milestones m on m.id = l.milestone_id
left outer join ac_projects p on p.id = l.project_id

Adding a custom field is easy; I do a mapping from the project name to determine the vertical and application.
select
  'insert into ticket_custom values (' +
  convert(varchar, t.id) + ', ' +
  dbo.strout('vertical') + ', ' +
  dbo.strout(
    case p.name
      when 'Project 1' then 'Vertical 1'
      when 'Project 2' then 'Vertical 2'
      when 'Project 3' then 'Vertical 1'
    end
  ) + ');',
  'insert into ticket_custom values (' +
  convert(varchar, t.id) + ', ' +
  dbo.strout('application') + ', ' +
  dbo.strout(
    case p.name
      when 'Project 1' then 'Application 1'
      when 'Project 2' then 'Application 2'
      when 'Project 3' then 'Application 1'
    end
  ) + ');'
from ac_project_tasks t
join ac_project_task_lists l on l.id = t.task_list_id
left outer join ac_project_milestones m on m.id = l.milestone_id
left outer join ac_projects p on p.id = l.project_id

And finally moving the comments over is straightforward.
select
  'insert into ticket_change values (' +
  convert(varchar, c.rel_object_id) + ', ' +
  convert(varchar, dbo.SQLite3_DateToJulian(c.created_on)) + ', ' +
  dbo.strout(dbo.acuser(c.created_by_id)) + ', ' +
  dbo.strout('comment') + ', ' +
  dbo.strout('1') + ', ' + -- this should really sequence for each comment for each ticket
  dbo.strout(c.text) + ');'
from ac_comments c
where rel_object_manager = 'ProjectTasks' -- not concerted about messages, only tasks
order by rel_object_id, created_on

There is a lot of cheating going on here but it is mostly for historical purposes. Just take the output of each section of routines and run them against the trac database.

Good luck!

I am able to capture the engine load at a point in time. What if I calculate an average load for a commonly traveled route? I can then take variations of that route and determine which path minimizes the impact to the car. I could factor a lot of variables including time, distance, average engine load, gas consumption, frustration, etc. I could then determine a set of heuristics that could suggest the optimal path.

This could be useful for frequently traveled routes like to and from work daily. Or any other profession that does a fair amount of repetitive automobile travel. I believe the software could just sit there gathering information with little input from the user. This would allow it to collect a vast amunt of info and make decent determintions based on user preferences.