So, here is my list of things which should to be done to get a project on the right track…most of them should have been taught in Project Management 101, but hey it seem that some people don’t know about them:

1. When a project is started…please tell the project manager that he/she is actually the project manager.
2. The project manager should ideally be involved in the project from the beginning and not just brought in when the shit hits the fan.
3. Tell the project manager what the main priority for the project is. Is it time, features, quality, people, cost…and NO it can’t be more than one…NO!!
4. Rank the features in the application according to priority, so that when push comes to shove, and it will, the PM knows what to concentrate on and what can be left to a later stage.
5. Please try to tell the project manager what his responsibilities and what the project involves
6. Don’t overrule the project managers decisions or undermine him…otherwise, you might as well hire a monkey from Primate Programming. Let the man do his job!!
7. Have clear communication channels about who should make decisions…a project should be run like a tight ship with clarity about these things. We can’t have a monkey jumping on the band wagon every time they want to make a change or don’t agree with something.
8. Listen to your technical staff about what they recommend and please take their advice on board…they do generally know what they are talking about – honest
9. Never override a developer estimate and never force him to do it quicker than he originally said he could…this will result in bad and buggy code.
10. If it could screw up, it will…oh yes it will, trust me!
11. Don’t do testing in the evening and yes you do need to hire a tester…
12. What constitutes project completion? What are the release criteria? And who will sign it off? Will that person be available to sign the project off, or will they be on holiday?
13. Get a decent PM, decent designer, a decent programmer, a decent sys admin…you need these people, so just hire someone good and pay them what they want.
14. Get a prototype out asap, so that you can prove the concept actually works
15. Don’t skimp on infrastructure to save costs on IDEs, servers etc. This saving will bite you in the bum at a later stage
16. And let us not forget…get the developers to use the source control in place…it’s there for a reason, so get them to use it damn it!

There are loads more points that could be added to this, but I think the post will become more of an essay then, so I’ve tried to include the most obvious ones. I have experienced most of these things and the question I keep asking is “What is this? Amateur hour?”. So, good luck to all the PM’s out there – Give it horns!

Thanks to Gareth and Drew for the banter – Nice one!

We have also released a new blog on blog.kindo.com, so if you would like to keep up to date with our latest news, then please visit it. Comments are more than welcome

Here are some snippets from two interesting posts on the Kindo blog

http://blog.kindo.com/2007/08/02/some-technical-details/en/

We’re having fun building the Kindo application and used a number of frameworks and components to achieve the finished result.

For the web framework we’re using Seagull which offers excellent integration with Flash through the AMF PHP request handling and easy integration with Ajax and Graphviz. For the Ajax interactions we use jquery exclusively, for Graphviz a heavily extended version of the PEAR lib of the same name. Working with the dot language has been interesting and a good opportunity to experiment with the subtleties of a genealogy graph.

http://blog.kindo.com/2007/09/04/welcome-to-the-beta-release/en/

A little while ago we released an Alpha to a small group of people, to gather valuable user feedback. Thanks to those great people, we’ve had our heads down improving Kindo, and as a result we’re releasing the Beta, and at the same time opening it up to a wider audience for more feedback.

The site which I am talking about is http://bd4d.com, which stands for By Designers 4 Designers is built using the seagull php framework with the new cms module

General comments

Media Temple Installation

During the live installation I encountered a few problems, most of which related to permissions of files on the server:

The main one however were the permissions of the /var/lib/php/session/ folder – it has to be writable

Plugins that I found really useful during development

There were two firefox plug-ins that I found very useful during development of the site:

Colorzilla – gives you # codes for colours on any webpage
IE tab – allows you to open up ie7 tabs within firefox, which is great for cross browser compliant xhtml/css

Thanks to the developers of these two add-ons

Paypal

The site has a payment gateway for listing jobs and events which has transaction processing done by Paypal.

In terms of project mistakes, this was one of the biggest we made. The support that James received from Paypal was extremely poor and the general documentation which they supply is often out of date and incorrect. I am amazed that they have become such a successful company, when it is so difficult to integrate their transaction processing into one’s site.

However after a lot of waiting for accounts to go live, numerous queries, emails, phone calls and a lot of frustration, we did eventually get the transactions working. It is working fine now, but it was definitely not worth all of that hassle.

I would be interested to here about other experiences from developers using Paypal – hopefully you had a better one than we did.

Site Features

Showcase

The section of the site that I am most proud of is the Showcase section, which has flash and movies of some really cool clips which the bd4d guys have done. The section allows comments to be made on each of the movies and I think it looks pretty slick – even if I say so myself

To get the section functioning, I used the following:

A comment manager which Dmitri initially developed and then I adapted that code to be used for the site.

To prevent spam on the site I used captcha package from the php pear library

And I used the following code in the template to hold the movies/flash files:

< object width="550" height="400" >
< param name="movie" value="{webRoot}/themes/bd4d/showcasemovies/
{oItem.media}" >
< param name="quality" value="high" >
< embed src="{webRoot}/themes/bd4d/showcasemovies/
{oItem.media}" quality="high" width="{oItem.width}" height="{oItem.height}" autostart="true">< / embed >
< / object >

Events

The Events calendar uses the default seagull events module as a basis, but there where a few enhancements made to fit the requirements of the site. The majority of the development of this section was done by James and I think that he did a pretty good job of it.

The calendar looks pretty slick and works really well, also the backend of the seagull events module provides a lot of flexibility for administrators wanting to manage events.

Job board

The Job board uses custom managers in the default module and as I have mentioned in my section about paypal, it allows members of the site to post jobs for a fee. The transaction processing handles straight paypal payments as well as credit card payments which are verified by paypal.

Conclusion

As I mentioned in the introduction this site took way longer than expected and there were a number of hard lessons which the whole team learnt. But at the end of the day, the site looks pretty good and works as expected, so in some way it can be seen as a success.

Background

I developed the site using the Seagull php framework, using the publisher module (which has recently been superseded by the CMS module).

The publisher module requires a bit of hacking to make it run effectively, especially if the site requires content pages which have different layouts. But, overall the site works well and looks pretty good, even if I say so myself .

I have still had a number of difficulties, so for those of you who are planning on building a similar site, here are some of the things which I have learnt:

This is especially relevant if you are going to be using some of the following:

• Seagull php Framework
• Installing a site on a Media Temple Server
• Including a dynamic Google Map in a page using the api

Lessons learned

Tips for using the Seagull publisher module:

If you have various site sections which need to have different layouts, the best way of handling this is to create different content types for each different layout which is required.

Then create a different publisher template for that content type and specify the styling which you want for that particular section.

If you have any exceptions, then use a handler like this:

{if:isEqual(29,articleID)}

{end:}
{if:isEqual(30,articleID)}

{end:}
{if:isEqual(34,articleID)}

{end:}

I will endeavour to write a more in depth tutorial for this and put it on the Seagull wiki. Once I have done so, I will update this post.

Server Issues

The first issue which I encountered was trying to do an installation on a staging sub-domain on a Media Temple Dedicated Virtual Server, where php safe mode was turned on. I have written a post which is a guide on how to successfully remove these restrictions.

Browser Issues

As always, I encountered problems with cross browser compatibility. I discovered huge differences between IE6, IE7 and Firefox. The site looks best in Firefox, Safari and in IE6, but there are a few small differences on IE7.

I used this basic conditional block to include an IE hacks file for when a user views the site using IE.

< !--[if IE] >< style type="text/css" >@import "http://www.atalink.com/ themes/atalink/css/hacks/ie.css";< /style >< ![endif]-- > (Remove spaces)

Google Maps

Although this is not that difficult to implement, I found quite a useful site called:

Google Maps latitude, longitude Popup , which gives you the latitudes and longitude values for a specific location, which you use in the Google supplied javascript

Design

Although a designer may do a really good job and a client signs off a design, one has to be flexible enough to deal with client requests for changing of the layouts. Often the changes requested are quite valid and one should consider each request before just disagreeing.

Content

Again, I have been reminded about how important it is to have content ready and signed off, before beginning with development. It is amazing how much time can be wasted when you are waiting for pieces of content. Details about content should be agreed upon during initial agreements and no matter how difficult it may be, these agreements should be adhered to.

Conclusion

I have enjoyed working on this site and I have learnt a lot, in terms of development and in terms of communicating with clients.

It is really easy to set the test suite up and to get going, all you really need are the following two downloads:

1. Selenium core test suite
2. To record tests the Selenium IDE in Firefox

Once you have installed the Selenium core web application and the plug-in for Firefox (Firefox rocks!!) you good to go:

1. Plan the various tests which you would like to run
2. Set the Firefox IDE to record your actions
3. Once you’ve completed a certain test save the test in an .html file
4. Include a link to your test file in your main index.html (TestSuite) file, which acts as a table of contents for the various tests
5. Then you can either run the tests individually or as a batch

If you would like a more detailed tutorial, I found the one on the thinkPHP blog to be quite useful.

As Demian says, you start getting “addicted” to creating tests and eventually try to become a perfectionist – which is a good place to be in if you want to produce error free sites for clients.

The problems I have encountered generally relate to project requirements and client expectations. What I have realised is that for a project to be successful and to be delivered on time (which hardly ever happens with IT projects) it is incredibly important to agree the requirements before starting.
Once these requirements have been agreed, there should be a line drawn in the sand. From that point no changes should be made to the original agreement.

Another thing that I have become aware of, which relates to expectations is that the average person wanting a webiste doesn’t care about technical specifications. As long as the site “works” then they are satisfied – They don’t care whether the site is table based and are more worried about the design then about fuctionality. This brings me to my “analogy”.

Whilst at lunch one day this week, we come up with “The house analogy” – lets call it that for now. What it entails is that I want to compare building a website to a house. When someone wants a house built they need to get a plan, blueprints, approval, building suppluies and other resources. What I am getting at is that there is a lot of planning reuired to build a house – you don’t just decide to build a house one day and then shoot from the hip.

I believe that the same should go for a website. I can’t get my head around the fact that people want a site developed without ever going through a process of planning. It is not sufficient to just provide a developer with some design layouts and then expect him to magically create a new website.

You can’t just give a picture of a house to a builder and say – “Build it, I want a new double story house with a balcony”. I am sure he would tell you to fly a kite! Well being a developer, it is so important to have proper requirments. I am talking about wireframes, process flow, sitemaps, information architecture and design. Not just a few pictures of a site.

The bottom line is that generally the average joe on the street wanting a webiste doesn’t want to spend extra money on that sort of planning detail and would rather go for the cheapest and quickest option. Sad but true!

All I can hope is that in the future, web standards will become more important and proper planning will become integral. This will make life as a develope lot easier!