Software application development
This blog is intended to provoke discussion and exchange between like minded software application developers, engineers, architects, project managers - and keen hobbyists too.
Friday 27 February 2009, 8:15 AM
IBM: Agile Is Not A Religion
Talking last year with Scott Ambler who is IBM’s global lead for Agile development at the company’s Rational Software Conference, I made the mistake of suggesting that the methodologies and processes behind Agile almost form a kind of religion for developers who follow its creed.
If you look back at my report last year, you’ll see that I tried to work this issue in. Ambler in fact told me in no uncertain terms that it is not a question of faith and I spoke to him again this week at length to find out what I should have said.
“Agile is a process and is sometimes thought of as a religion, but this just gets in the way in my opinion – I want to get past the rhetoric and get down to fact-based arguments,” said Ambler.
… and I tell you, Ambler is a really relaxed and laid back guy – but he’s straight down the line with this one and he’s not prepared to mess around.
IBM’s Scott Ambler
We actually started having this discussion on the basis of how SCM dovetails with Agile if it needs to scale. The argument being that SCM becomes even more important in Agile because of the large amount of shared artefacts. So as Agile works in a very iterative manner it’s critical to be able to lock this stuff and feed into and out of a central code repository.
“To really scale Agile you need to address architecture issues and think about scaling from the start as you strategically break up teams. You need to be doing systems integration between these teams on a regular basis – and there needs to be a major focus on teamwork and collaboration,” said Ambler.
Well, we know IBM can’t say the words ‘teamwork and collaboration’ enough. It has themed it’s last four global developer conferences around team structure and integration.
“In terms of scaling requirements – there is no such thing as a really large team, there are only teams made up of sub teams, each looking after building their own sub systems. Each sub system will have its own requirements and cross team requirements must not clash in terms of their dependencies on the central code base - NOT only this, but they must also all work in tandem with the overall macro scale requirements of the entire project. Plus, of course, these requirements will always change over time,” said Ambler.
To be honest with you, we had this discussion outside of the context of any press release or conference of any shape – and that made it so much of a better briefing. Religion or not, I was impressed by Ambler’s refusal to adopt a cute branding tag to promote his technology and insist upon a resolute focus on facts – and just the facts.
It’s not all good news for Agile fans (whoops – almost said disciples!) though and as I was on the hunt for Agile opinions I spoke to analyst Jon Collins from Freeform Dynamics who pointed me to an opinion he had stated previously.
According to Collins, “Agile approaches are not just something you can pick up and run with. There is little room for error: timescales are divided into variously named short periods (e.g. ‘sprints’, or ‘timeboxes’), which are kicked off with a prioritisation exercise as to what is to be delivered and finished off with a delivery. If something goes wrong, there can be a domino effect on other parts of the project. To all intents and purposes, Agile approaches can be intense and rewarding, but the one thing they are reliant on is a level of structure and co-ordination.”
For a nice leveling statement I also spoke late on Thursday to Dave Robertson who is VP of International for SCM specialists Perforce. He told me, “For all the recent excitement on the subject, the principle values of Agile software development were enshrined by a number of leading software engineers in the ‘Manifesto for Agile Software Development’ back in early 2001. Crucial to these values are that individuals and interactions are more important than the processes and tools being used.”
… and guess what? Borland wanted to send me an opinion too, so leaving out the lovely bits of news they sent me about their latest awards, global rankings, speed and price, here’s the technology comment freshly prepared for ZDNet.co.uk…
“Borland has transitioned 65% of our 300-person distributed development organisation to Agile. As a result, we have achieved a 100% increase in product releases per year; reducing ‘open’ issues from release to release by 50% and streamlined reporting and project administration. We’re not saying Agile is easy or without its challenges – the transformation process takes time and commitment in order to achieve results. Yet, done properly, an ‘Agile Shift’ can revolutionise the way a business operates, not just in software development or IT, but across the wider organisation,” said Andy Seager, global vice president of product strategy at Borland.
So there you have it. Don’t people get passionate when they talk about Agile though? It’s good isn’t it? I mean, I wish we could get this fired up about every aspect of software development.
I was on the phone to a guy from a hosting company yesterday and he told me that his software looked after the automated billing function for virtualised services and that it wasn’t the sexiest aspect of the technology offering from his company. I told him it was – it’s the commercial front end!
Anyway, don’t they say to never talk about sex, politics or religion if you want to avoid an argument? Perhaps we should bring all three into technology reporting to stoke up the fire then!
If you look back at my report last year, you’ll see that I tried to work this issue in. Ambler in fact told me in no uncertain terms that it is not a question of faith and I spoke to him again this week at length to find out what I should have said.
“Agile is a process and is sometimes thought of as a religion, but this just gets in the way in my opinion – I want to get past the rhetoric and get down to fact-based arguments,” said Ambler.
… and I tell you, Ambler is a really relaxed and laid back guy – but he’s straight down the line with this one and he’s not prepared to mess around.
IBM’s Scott Ambler
We actually started having this discussion on the basis of how SCM dovetails with Agile if it needs to scale. The argument being that SCM becomes even more important in Agile because of the large amount of shared artefacts. So as Agile works in a very iterative manner it’s critical to be able to lock this stuff and feed into and out of a central code repository.
“To really scale Agile you need to address architecture issues and think about scaling from the start as you strategically break up teams. You need to be doing systems integration between these teams on a regular basis – and there needs to be a major focus on teamwork and collaboration,” said Ambler.
Well, we know IBM can’t say the words ‘teamwork and collaboration’ enough. It has themed it’s last four global developer conferences around team structure and integration.
“In terms of scaling requirements – there is no such thing as a really large team, there are only teams made up of sub teams, each looking after building their own sub systems. Each sub system will have its own requirements and cross team requirements must not clash in terms of their dependencies on the central code base - NOT only this, but they must also all work in tandem with the overall macro scale requirements of the entire project. Plus, of course, these requirements will always change over time,” said Ambler.
To be honest with you, we had this discussion outside of the context of any press release or conference of any shape – and that made it so much of a better briefing. Religion or not, I was impressed by Ambler’s refusal to adopt a cute branding tag to promote his technology and insist upon a resolute focus on facts – and just the facts.
It’s not all good news for Agile fans (whoops – almost said disciples!) though and as I was on the hunt for Agile opinions I spoke to analyst Jon Collins from Freeform Dynamics who pointed me to an opinion he had stated previously.
According to Collins, “Agile approaches are not just something you can pick up and run with. There is little room for error: timescales are divided into variously named short periods (e.g. ‘sprints’, or ‘timeboxes’), which are kicked off with a prioritisation exercise as to what is to be delivered and finished off with a delivery. If something goes wrong, there can be a domino effect on other parts of the project. To all intents and purposes, Agile approaches can be intense and rewarding, but the one thing they are reliant on is a level of structure and co-ordination.”
For a nice leveling statement I also spoke late on Thursday to Dave Robertson who is VP of International for SCM specialists Perforce. He told me, “For all the recent excitement on the subject, the principle values of Agile software development were enshrined by a number of leading software engineers in the ‘Manifesto for Agile Software Development’ back in early 2001. Crucial to these values are that individuals and interactions are more important than the processes and tools being used.”
… and guess what? Borland wanted to send me an opinion too, so leaving out the lovely bits of news they sent me about their latest awards, global rankings, speed and price, here’s the technology comment freshly prepared for ZDNet.co.uk…
“Borland has transitioned 65% of our 300-person distributed development organisation to Agile. As a result, we have achieved a 100% increase in product releases per year; reducing ‘open’ issues from release to release by 50% and streamlined reporting and project administration. We’re not saying Agile is easy or without its challenges – the transformation process takes time and commitment in order to achieve results. Yet, done properly, an ‘Agile Shift’ can revolutionise the way a business operates, not just in software development or IT, but across the wider organisation,” said Andy Seager, global vice president of product strategy at Borland.
So there you have it. Don’t people get passionate when they talk about Agile though? It’s good isn’t it? I mean, I wish we could get this fired up about every aspect of software development.
I was on the phone to a guy from a hosting company yesterday and he told me that his software looked after the automated billing function for virtualised services and that it wasn’t the sexiest aspect of the technology offering from his company. I told him it was – it’s the commercial front end!
Anyway, don’t they say to never talk about sex, politics or religion if you want to avoid an argument? Perhaps we should bring all three into technology reporting to stoke up the fire then!
Tuesday 24 February 2009, 8:05 AM
It’s Virtual Reality Time in the Data Centre
With VMworld Europe kicking off this morning in Cannes – and the hullabaloo over the Microsoft Hypervisor situation mercifully now behind us – it’s anybody’s guess as to what will make the headlines this year.
One can only imagine that the ‘reality’ factor will come to the fore now as virtualisation starts to seep into the corporate blood stream of IT stacks from Cannes to Calcutta. One would imagine that a large part of that reality factor may be related to the ‘touch point’ at which the average user comes into contact with virtualised technology i.e. applications.
Virtualised desktop applications (or VDI virtualised desktop infrastructure as Parallels likes to call it) will now need some element of control, or dare I say it – management, if we are to know who is using what and where they are using it – and, crucially for virtualised apps, if they are using it ‘optimally’.
I know that sounds slightly cheesy, but surely optimal performance is the area that virtualised app administrators will want to look at most if the they are to balance the deployment of virtual machines based on real-time measurements performance.
Chris Wolf, Burton Group senior analyst said recently, “There’s urgency in enterprises today to deploy virtualisation to improve server utilisation and reduce overall IT expenditures. However, broad deployment of virtualisation in business-critical applications has been limited by poor visibility into how the SAN I/O infrastructure affects the performance of applications deployed on virtual machines. The only way to overcome that hurdle is to implement instrumentation and measurement tools within the SAN that extract real-time transaction performance information.”
Now I couldn’t find out whether they are at VMworld Europe 2009 as I am lamentably not able to attend this year. But I did find a company in this space that talks on (and logically also sells) performance analysis solutions. Virtual Instruments has a tool-kit that claims to provide real-time and historical insights into application latency and bandwidth consumption.
I do them a disservice – I’m sure it’s more than a tool-kit, but this is a quick blog and not an analysis. Their VirtualWisdom product is said to be customisable widget-based dashboard for sys admins to focus on specific virtual machine, server and application I/O data points that matter most to them. You can read more on their web site, which I’m sure you are capable of finding.
Company CEO Mark Urdahl seems to have it right though when he says that 2009 will be all about looking at the investment made into virtualised infrastructures, examining the performance of the applications within the virtual stack and, logically, achieving higher server consolidation ratios as a result.
As for VMworld, ah it’s webinars and media alerts for me this year. But I still put the two-hour French breakfast I had on the Sunday before last year’s event as one of my top ten meals of all time. Just for the record – you know that thing they do in fancy hotels where they put a bottle of fizz out for you to have maybe a half glass with your brekky brochettes? Well, I absolutely did not drink the whole bottle and ask for another OK? It was half empty to start with and I had help.
One can only imagine that the ‘reality’ factor will come to the fore now as virtualisation starts to seep into the corporate blood stream of IT stacks from Cannes to Calcutta. One would imagine that a large part of that reality factor may be related to the ‘touch point’ at which the average user comes into contact with virtualised technology i.e. applications.
Virtualised desktop applications (or VDI virtualised desktop infrastructure as Parallels likes to call it) will now need some element of control, or dare I say it – management, if we are to know who is using what and where they are using it – and, crucially for virtualised apps, if they are using it ‘optimally’.
I know that sounds slightly cheesy, but surely optimal performance is the area that virtualised app administrators will want to look at most if the they are to balance the deployment of virtual machines based on real-time measurements performance.
Chris Wolf, Burton Group senior analyst said recently, “There’s urgency in enterprises today to deploy virtualisation to improve server utilisation and reduce overall IT expenditures. However, broad deployment of virtualisation in business-critical applications has been limited by poor visibility into how the SAN I/O infrastructure affects the performance of applications deployed on virtual machines. The only way to overcome that hurdle is to implement instrumentation and measurement tools within the SAN that extract real-time transaction performance information.”
Now I couldn’t find out whether they are at VMworld Europe 2009 as I am lamentably not able to attend this year. But I did find a company in this space that talks on (and logically also sells) performance analysis solutions. Virtual Instruments has a tool-kit that claims to provide real-time and historical insights into application latency and bandwidth consumption.
I do them a disservice – I’m sure it’s more than a tool-kit, but this is a quick blog and not an analysis. Their VirtualWisdom product is said to be customisable widget-based dashboard for sys admins to focus on specific virtual machine, server and application I/O data points that matter most to them. You can read more on their web site, which I’m sure you are capable of finding.
Company CEO Mark Urdahl seems to have it right though when he says that 2009 will be all about looking at the investment made into virtualised infrastructures, examining the performance of the applications within the virtual stack and, logically, achieving higher server consolidation ratios as a result.
As for VMworld, ah it’s webinars and media alerts for me this year. But I still put the two-hour French breakfast I had on the Sunday before last year’s event as one of my top ten meals of all time. Just for the record – you know that thing they do in fancy hotels where they put a bottle of fizz out for you to have maybe a half glass with your brekky brochettes? Well, I absolutely did not drink the whole bottle and ask for another OK? It was half empty to start with and I had help.
Monday 23 February 2009, 9:34 AM
Is Column-based Relational Analytics a Database Cash Cow?
I spent last Thursday night getting a pre-briefing on the launch of the Sybase IQ v15 analytics server with the company’s VP of engineering Richard Pledereder, who gave me some pointers as to why this type of data manipulation ‘engine’ is proving to be such a revenue booster during the global economic slowdown.
To set the scene - if you look at the world economy through a wide lens, you may be able to pinpoint a few areas where demand will always be high. I would like to argue that Internet usage, mobile SMS texts, the financial services industry and defence are all perennial must-haves.
All good areas for database companies to make some cash I think.
Pledereder pointed out to me that now there is greater corporate pressure to increase risk management provisioning for data, as well as increased compliance pressure all round. He also said that column-based relational analytics has gained wider recognition as a process for managing data, especially if it is subsequently fed into business intelligence applications.
If this is a good message for developers who may be working on business intelligence apps that exist within data-centric environments where the number of transactions and the number of users is exploding, then I hope I am doing my job by talking about it.
This strain of enterprise analytics for databases (as it touches both DBAs and developers alike) is argued to be well suited to work with real-time data environments, especially inside data shops where the project manager has a particular need to be able to use the analytics to track customer preferences and adjust business processes quickly.
While the rest of you were watching Coldplay not win anything at the Brit awards, I was on the phone to Dublin (the one in California) discussing the above subject. Pledereder was obviously full of the expected corporate messages, so let’s gloss over favourites like speed, flexibility, manageability and ROI and suggest that the lower load latency for large data loads capability stands out as a newer area of interest.
The announcement of this new version of Sybase’s product today is, I was told, the biggest overhaul of the product in the last five to seven years. As I have attended the company’s TechWave developer and DBA convention once or twice over the same time period I thought it would be appropriate to mention it here.
I was at a lunch with Sybase CEO John Chen last year (OK, there were 40 other people in the room too) and he was kind of self-effacingly modest about business. “Bank on us,” he said – urging the room to think of his company as worthy competitors for Oracle and IBM. It seems like a few parties out there do.
To set the scene - if you look at the world economy through a wide lens, you may be able to pinpoint a few areas where demand will always be high. I would like to argue that Internet usage, mobile SMS texts, the financial services industry and defence are all perennial must-haves.
All good areas for database companies to make some cash I think.
Pledereder pointed out to me that now there is greater corporate pressure to increase risk management provisioning for data, as well as increased compliance pressure all round. He also said that column-based relational analytics has gained wider recognition as a process for managing data, especially if it is subsequently fed into business intelligence applications.
If this is a good message for developers who may be working on business intelligence apps that exist within data-centric environments where the number of transactions and the number of users is exploding, then I hope I am doing my job by talking about it.
This strain of enterprise analytics for databases (as it touches both DBAs and developers alike) is argued to be well suited to work with real-time data environments, especially inside data shops where the project manager has a particular need to be able to use the analytics to track customer preferences and adjust business processes quickly.
While the rest of you were watching Coldplay not win anything at the Brit awards, I was on the phone to Dublin (the one in California) discussing the above subject. Pledereder was obviously full of the expected corporate messages, so let’s gloss over favourites like speed, flexibility, manageability and ROI and suggest that the lower load latency for large data loads capability stands out as a newer area of interest.
The announcement of this new version of Sybase’s product today is, I was told, the biggest overhaul of the product in the last five to seven years. As I have attended the company’s TechWave developer and DBA convention once or twice over the same time period I thought it would be appropriate to mention it here.
I was at a lunch with Sybase CEO John Chen last year (OK, there were 40 other people in the room too) and he was kind of self-effacingly modest about business. “Bank on us,” he said – urging the room to think of his company as worthy competitors for Oracle and IBM. It seems like a few parties out there do.
Friday 20 February 2009, 9:07 AM
Forms-based Applications & the Web 2.0 ‘Human’ Factor
I often wonder if the exasperation some of us experience with call centres, utility companies and even our banks is mainly caused by poorly provisioned IT systems, or poor service caused by operator errors.
My own experience with my car insurance company make quite a tale – but it’s a story with a happy ending where technology came to my rescue in the one of the coolest ways I’ve ever experienced.
In short, Norwich Union (soon to be renamed Aviva) and I have a very close relationship after the hours (plural) I spent on the phone to Bangalore (soon to be renamed Bengaluru) getting my wife (recently renamed Mrs Bridgwater) some insurance organised.
Incompatibilities between different insurer’s systems can make this task a struggle. Some have a field to tick for maximum no claims being four years and some for five. It’s worse still if you are American and you have to try and explain what a GEICO ‘Good Driver’ discount is.
When the software is built for these call centre operatives to use – do the companies open up and look at the systems of their competitors? I would hazard a guess not - right? Forms-based applications that must align across different competitors within the same industry could and should benefit hugely if they had better sight of each other – right?
Suffice it to say that I was pulling my hair out (yes I have a bit left) and started putting out emails and Tweets to say how angry I was. Now, here’s the good bit…
Norwich Union employs an “online customer experience consultant” person for the UK. Let’s call her Becca Sibley (as that’s her name) – who was introduced to me via a diligent PR contact who had been watching me rant.
Becca not only fixed the problems immediately, she also provided me with compensation for my troubles and turned me from a very angry customer into an extremely happy one.
Is this where social networking applications can really start to plug into our daily interaction with seemingly hidden enterprise apps housed within the data vaults of the companies that we have to depend upon for essential services? People are talking every now about Twitter going mainstream in 2009. So was Becca’s service a perfect example of how we’ll now use these technologies.
Here’s my argument: build a behemoth of a corporate IT application stack, slap on a few databases, bring in web interfaces and associated services, promise to deliver ROI at xyz level and it’ll never go wrong, ever.
Well, given the human factor (and I’m not going to blame the Indian call centres as they tried hard with my slightly more complex than normal requests) – things are always going to go wrong now and again.
Seeking an opinion on this subject I spoke directly to a chap called Tilman Eberle who is VP of communications at a company called Doodle. Tilman told me, “One of the big reasons web 2.0 tools grow virally, so quickly, is that they can put the ‘human’ back into customer relations. They give companies a face and real people start interacting, not just systems. For instance, if you have a problem with your Twitter account, you can send a customer service request to an actual Twitter person by simply @’ing them. This type of real-time, interactive, 2.0-based customer relations could become more popular.”
So if enterprise apps need to serve customers effectively, perhaps they need a ‘human’ Internet-driven front end like Becca to pick up on anomalies and ensure the company’s reputation remains intact.
Is this one for your project manager or system architect to think about in 2009 and beyond? I hope so.
My own experience with my car insurance company make quite a tale – but it’s a story with a happy ending where technology came to my rescue in the one of the coolest ways I’ve ever experienced.
In short, Norwich Union (soon to be renamed Aviva) and I have a very close relationship after the hours (plural) I spent on the phone to Bangalore (soon to be renamed Bengaluru) getting my wife (recently renamed Mrs Bridgwater) some insurance organised.
Incompatibilities between different insurer’s systems can make this task a struggle. Some have a field to tick for maximum no claims being four years and some for five. It’s worse still if you are American and you have to try and explain what a GEICO ‘Good Driver’ discount is.
When the software is built for these call centre operatives to use – do the companies open up and look at the systems of their competitors? I would hazard a guess not - right? Forms-based applications that must align across different competitors within the same industry could and should benefit hugely if they had better sight of each other – right?
Suffice it to say that I was pulling my hair out (yes I have a bit left) and started putting out emails and Tweets to say how angry I was. Now, here’s the good bit…
Norwich Union employs an “online customer experience consultant” person for the UK. Let’s call her Becca Sibley (as that’s her name) – who was introduced to me via a diligent PR contact who had been watching me rant.
Becca not only fixed the problems immediately, she also provided me with compensation for my troubles and turned me from a very angry customer into an extremely happy one.
Is this where social networking applications can really start to plug into our daily interaction with seemingly hidden enterprise apps housed within the data vaults of the companies that we have to depend upon for essential services? People are talking every now about Twitter going mainstream in 2009. So was Becca’s service a perfect example of how we’ll now use these technologies.
Here’s my argument: build a behemoth of a corporate IT application stack, slap on a few databases, bring in web interfaces and associated services, promise to deliver ROI at xyz level and it’ll never go wrong, ever.
Well, given the human factor (and I’m not going to blame the Indian call centres as they tried hard with my slightly more complex than normal requests) – things are always going to go wrong now and again.
Seeking an opinion on this subject I spoke directly to a chap called Tilman Eberle who is VP of communications at a company called Doodle. Tilman told me, “One of the big reasons web 2.0 tools grow virally, so quickly, is that they can put the ‘human’ back into customer relations. They give companies a face and real people start interacting, not just systems. For instance, if you have a problem with your Twitter account, you can send a customer service request to an actual Twitter person by simply @’ing them. This type of real-time, interactive, 2.0-based customer relations could become more popular.”
So if enterprise apps need to serve customers effectively, perhaps they need a ‘human’ Internet-driven front end like Becca to pick up on anomalies and ensure the company’s reputation remains intact.
Is this one for your project manager or system architect to think about in 2009 and beyond? I hope so.
Wednesday 18 February 2009, 8:26 AM
Application architecture data published from 2,500 open source projects
Having just finished working on a 10-page tutorial detailing the inner workings of the tabular data stream protocol in relation to database administration and related gateways for programmers, I think its safe to say that I realise developers like to get down to a granular level when it comes to code and the examination (and perhaps the ‘classification’) if it.
With this is mind; I mention the fact that that this week sees the publication of application architecture data from over 2,500 popular Open Source software projects.
This coder’s nomenclature (if you will) is a result of an extension of a contract with the U.S. Department of Homeland Security with a company called Coverity that says it collected this data using its own analysis tools.
This publicly accessible resource can be accessed here. It includes application architecture files and diagrams designed for developers planning to incorporate open source packages in their applications.
It is also billed as a resource for developers that want to learn the architectures of successful projects to improve the structure of their own applications.
This library is a database of application architecture diagrams from open source projects such as Amanda, NTP, OpenPAM, OpenVPN, Overdose, Perl, PHP, Postfix, Python, Samba and TCL.
It was created using Coverity Architecture Analyzer product which claims to be able to automatically map the relationships between code elements at the function and file levels, identifying the underlying structure of software to help developers identify violations of architectural standards.
According to Coverity, “The ability to study a visual presentation of an application’s architecture and related data offers a number of benefits to developers. For example, developers planning to use or build on top of a project in the Scan library can optimise their use of it by comparing architecture plans to other codebases that use the same project.”
Sounds like it’s worth a click I would venture. My better half and I are in the midst of building new web sites and projects that should necessitate the resurrection of “that old laptop in the loft” to work as a web server.
Now we’ll probably just opt for Apache, but if we were building our own web server, wouldn’t it be cool to be able to study the architecture of open source projects to improve the development of our own application architecture?
With this is mind; I mention the fact that that this week sees the publication of application architecture data from over 2,500 popular Open Source software projects.
This coder’s nomenclature (if you will) is a result of an extension of a contract with the U.S. Department of Homeland Security with a company called Coverity that says it collected this data using its own analysis tools.
This publicly accessible resource can be accessed here. It includes application architecture files and diagrams designed for developers planning to incorporate open source packages in their applications.
It is also billed as a resource for developers that want to learn the architectures of successful projects to improve the structure of their own applications.
This library is a database of application architecture diagrams from open source projects such as Amanda, NTP, OpenPAM, OpenVPN, Overdose, Perl, PHP, Postfix, Python, Samba and TCL.
It was created using Coverity Architecture Analyzer product which claims to be able to automatically map the relationships between code elements at the function and file levels, identifying the underlying structure of software to help developers identify violations of architectural standards.
According to Coverity, “The ability to study a visual presentation of an application’s architecture and related data offers a number of benefits to developers. For example, developers planning to use or build on top of a project in the Scan library can optimise their use of it by comparing architecture plans to other codebases that use the same project.”
Sounds like it’s worth a click I would venture. My better half and I are in the midst of building new web sites and projects that should necessitate the resurrection of “that old laptop in the loft” to work as a web server.
Now we’ll probably just opt for Apache, but if we were building our own web server, wouldn’t it be cool to be able to study the architecture of open source projects to improve the development of our own application architecture?
Previous
