Pokazywanie postów oznaczonych etykietą architecture. Pokaż wszystkie posty
Pokazywanie postów oznaczonych etykietą architecture. Pokaż wszystkie posty

piątek, 21 marca 2008

Microsoft Solution Framework - tips and tricks

I have refreshed lastly my knowledge about the MSF. I have looked backward and deep into my soul and here there is a short list of my private suggestions if you think about going into the direction...

Choose carefully MSF3 or MSF4
There is quite a big twist in Microsoft approach to the MSF lastly. When MSF3 is the loose set of good practices written on paper, MSF4 is their implementation within VSTS... some people say even that MSF has been monetized. MSF4 defines two main templates and guidance - MSF for Agile and for CMMI. There is an option to define your own process with Template Editor, but that is not an easy process.
In general...

  • Option 1 - If you have small team and you want easily customize the process - MSF3 will be a good choice and set of VS Pro licenses will be fine.
  • Option 2 - If you work within big corporation, in distributed geographically team, keeping a standard is a problem... you may think about MSF4 and convincing stakeholders to spend a lot of $ for VSTS installation (that is not just about buying licenses - you have servers, administrators, infrastructure - all the heave equipment).

BTW, if you are in second option and you works with open technology, you need to consider RUP and Eclipse from IBM. There funny thing to write but in first option you may do .Net projects with RUP and you may do J2EE projects with MSF, even when MSF+Visual Studio belongs to Microsoft and RUP+Eclipse belongs to IBM.

Do not be religious
You just read what MSF is about, you are very excited... Lets rock, lets put it into play...
No, no... Hold on for a sec... go for vacation... Chill down...
Whatever you do, does not do it because the paper says so. MSF is just the tool to realize the goals. Which goals? What do you want to achieve? Whatever you do, does not forget that you are about improving things.

Introduce Framework step-by-step
Anayze the current status! Do you have some well working processes? Take care about them and protect them. Do not hesitate to customize MSF, so old-well-working things stays where they are.

If something works wrong, indentify 1-3 the biggest problems and solve them using particular mechanisms from MSF. Daily builds, bug triage, zero bug bounce - these are the good candidates at the beginning... Does it work? Do people like it? It is MSF... Lets start the new project with 5 phases, but if there is project manager in Prince2 meaning, do not hestiate to unite Program and Product manager role if that is necessary.

Team model in practise
Team model in MSF is quite unique and white paper says that each role has to work in each phase. That is not totally true. Release Management comes into play rather in the final phases. Keep in touch with the guy, but does not insist on having him at each meeting and 100% engaged into the project at any stage.

The other story is User Experience fellow. The main problem is, that there is so little people with this profile on the market. In practice very often the role is taken by business analist... and that is fine. In similar way the Program Manager role is often classic Project manager and Product Manager is Sponsor of the project. As far as it works... lets leave it ;)

Use trade-offs
Trade off theory has helped me often in practise to leverage new requirements and staying assertive - nothing is for free!

- Of course we can add the functionality! What do I get in return? Extra time or person? When he will be on the board? In 2 months? Ok... When he is on board we will analyze the new functionality ;)

In very taugh times you may mention fourth dimension - quality; triangle story extended to tetrahedron (three sided piramid)...

- I may agree to add new functionality, within the same timeframe and resources, but you must approve to me on paper that you accept lower quality of the final product.

... Yes, that is just the intelectual trip ;)

>Keep the balance between the phases
If envisioning took 1 week, the planning will probably take around 1 month. If during planning you have done all the key Proof of Concepts in code, you may close the developing phase in next 1 month. Stabilizing will the next 1 month and Deployment about 1 week.

For example... It is quite common error, that when things goes wrong the stabilization time is cut off. The effect? The same time doubled or tripled is spent during Deployment. That is a dark side of Project Management that in crasis, the taught situation must be faced and not just postponed.

If you prepare the schedule in planning phase, do not expect that you will be able to close 3 next phases in 2 months and 1 week, if getting to the point since the beginnig of the project took you over 2 months - there are some exceptional situation, but at least look in the mirror and answer yourself where these exceptions truely are.

Use bug triage
I bet my surname domain on the short setence - BUG TRIAGE WORKS!

Things, which are not monitored once per a time, go into the chaos stage. Let's face it - reparing bugs is boring and painful work to do for developers. Everybody knows it. Anyway it must be done though, in the same way how you must clean your flat (at least once per a while). Project Manager must be sometimes a mommy who says: "Sweetheart... Clean your room tommorow, ok?" or "When finally you will put these dirty socks into this damm wardrobe!" :D

Having whichever Bug Tracking System (excel / customized open source project / customized SharePoint poratl with lists / commercial product / VSTS ) is at great help - anybody may quickly see what happens with bug #645.

- Hey Bart! Our CEO has looked today in our BTS and he has asked about this problem with dockable windows. Can you add him some comment about #645 and send him a link?

...and this is it - you definitely will not loos by knowing what MSF is about ;)

poniedziałek, 7 stycznia 2008

UML, RUP, MSF and the others versus reality

Mood: Monday was not as horrible as expected
Link: I am not the fan of the serie, but the sounds keep pinging my mind ;)

"We need simply to use UML and it will fix all the problems" - I was quite surprised to see the idea in MY mind and luckily not in my mouth yet ;) I am quite sure you know also the people, who believes that particular approach like Prince2 or RUP will be the remedy for any diseas. The problem is that effectiveness of these solutions always depend on particular situation. That is always the most important to diagnose the problem first and then to find the medicine in various books ;) Sounds like the cliché, but there is still so many werewolf hunters fully equipped with silver bullets. I have made conscience-searching and here we are...

General
· None of these specify necessity of designing GUI, but usually when you start drawing how particular screens will look like you may gain a lot of information (unless you plan the prototype first or strictly agile, iterative approach)
· None of these (but Prince2 books mention it unofficially) specify necessity of hard-numbered profit versus cost estimation. Business analysis should include some predictions of key indicators like Net Present Value or Internal Rate of Return
· QA is different topic, which you need to have in back of your head regardless from everything else (RUP and Prince2 mention it just)
· No one below will remind you about the law issues and regulations (legal acts, being compliant with standards, corporate regulations)

UML
· UML diagrams are really cool, but THAT IS NOT THE METHODOLOGY – that is simply some set of boxes, which you may use in drawings. Nothing more, nothing less.
· Very many companies uses UML to attach some images to documentation (good), but unfortunately not too many generates even the classes skeleton from it; even less keep synchronization between the physical and logical model. Surprisingly I have never heard or read about any CMMI-like model, which would suggest it or score it!
· When we speak about Use Case diagram remember about drawing the border of the system and if necessary about the version of the system (quite often technical spec says about the subset of functionalities enlisted in business spec)
· Use Cases are also about identifying the type of users
· Apply packages to simplify Use Cases
· Always look for reuse of diagrams and avoid copy&pasting (eg. <>, <> in use cases)
· Think about Deployment Diagram and general plan on early stage; it is also a good moment to think which OSes and browsers might be used by final users (it is often forgotten and it blows out when the deployment truely starts)
· Sometimes it is quicker to scetch, scan and paste even sligthly against the rules, that draw the diagram with 100k EUR worth software beign strict; pretty-good is usually enough ;)
· UML is one of the best for desiging software, but it is not the only one - have a look here.

RUP
· That is quite cool approach, which nicely show 4 important stages of the project – preparation, analysis, realization and deployment, but there is no point where all the analysis is finished. It means that even when you start the 3rd Construction phase you do not have all use cases finished (just 80% is required)!
· Against RUP I prefer rather to start coding when I have complete set of uses cases and the range of first release fully established - I found too often one, nasty case, which caused total redesign
· Surprisingly RUP does not go deep within technical aspects - it mentions mostly use cases, business aspects and risk management.
· Martin Fowler in his book The New methodology says something like: “My experiences with RUP are, that you can customize it without the boundaries and it causes problems. I have met couple of RUP usages, starting from the cascade model with analytical iterations, finishing at the full Agile process. I was surprised that this promoting RUP as one process caused that people may do everything and call it RUP – it makes RUP the word without meaning” - I guess this happens because IBM strategy is to allow in the same time the marriage of it with SCRUM and Prince2
· You can quicker find interesting documentation about RUP outside IBM (like Wikipedia) than inside, where you must go through tons of commercial crap to get kilo of knowledge. The kilo is probably somwhere within 270 pages long red book.
· RUP defines couples of disciplines - among others Business Modelling, Requirements, Analysis & design. In situation, when you need to make the interview with clients (in order to specify the contract), it means that each specialty must be represented at each meeting.

MSF
· Unfortunately since the new version (at least the document is ONLY 47 pages long ;) ), the split between logical and physical design is not bolded so strongly. You need to spend significant time to find the sentence like “There are three levels in the design process: conceptual design, logical design, and physical design”. I guess that is caused by the necessity of generalization caused by MSF for Agile.
· Good trade-off mechanism, which finally learned how to be agile, when customer wants something more :D
· User experience role, which you will not find in RUP.
· Practical bottom-up approach for estimation and no word about it in RUP

I was thinking for a sec to weight both RUP and MSF and put the score, but the truth is that the covarage which both companies for their development environment (IBM with Eclipse and Microsoft with VS.Net). There is couple of interesting places, where you can find deeper comparsion of both like even some Master Tesis prefering slightly RUP. At the end of the road, it does not matter what label you wear, but how you feel with it. Is it flexible? Can you show in the suite on the business meeting? Can you run to catch the cab? Do you buy clothes in regular market or you can afford the tailor? Do not you think you need to have couple of clothes? :D

piątek, 30 listopada 2007

CardSpace xor OpenId - trends

Mood: My skis are packed, holidays incoming!

About CardSpace
CardSpace (formerly “InfoCard”) is the solution introduced within .Net Framework 3.0 side by:
- Windows Workflow Foundation (WWF)
- Windows Communication Foundation (WCF)
- Windows Presentation Foundation (WPF)
It is Microsoft’s standard for authentication and digital information transport. MS did in here 180 degree turn. Predecessor, Microsoft Passport was central type of the solution, where Microsoft severs possessed all the information about authenticated users. It did not become popular outside Redmond servers. Now, the user is the one who possess the information and the one who decides which information (Information Card) from his personal PC are passed to which web site. CardSpace is so the distributed solution, where MS provides just the standard of exchanging data and it has much bigger chance of success.

How it works, it is described in details on many, various pages to name just the one.

Fig1 – CardSpace model
If you know polish, you should also read this.

Personal Information Card – first steps
CardSpace defines two types of Information Cards:
- Personal Cards, which contains the standard, defined by MS set of information
- Managed Information card (use the package if you want to play with them)
At the beginning it is suggested to play with Personal Cards and install your personal one like it is described in here. Important note is though that you do not need Vista and it is working also for XP SP2 and Windows 2003 SP1.

You may register your card in WindowsLive or MyOpenId. MyOpenId is the interesting authentication mid-service for number of other web portals controlling your “sign in” action there with one assign id. After you sign in on these services for the first time, you are transferred to MyOpenId page and you may decide about your nickname there plus you are asked if you want to allow the action forever, once or deny it.

Fig2 – MyOpenId service

Nevertheless, after registration in both services you may transparently get to WindowsLive service, but MyOpenId will require from you sending your Personal Information Card during each Sign In action! The second situation is probably due to other model and high security level as any changes in MyOpenId affects the way of accessing number of other.

After these and the other actions you may always check usage history of your card.

Fig3 – History of Personal Information Card

CardSpace enabled Web page
When you want to create your own web page, which is CardSpace enabled (like MyOpenId) there is couple of tricks like:
- You need to create and install your some “X” High Assurance certificate issued by “Y” CA to enabled SSL (*.pfx file)
- Add “Y” CA certificate into list of Trusted Root Certificate Authorities (*.sst file)
- Specific page for choosing, displaying and registration the Personal Information Card
- Authentication mechanism for your page
All these things are described in details in couple of places, just to name one.

Further more Fireworks 3.0 (likely) supports CardSpace.

OpenID
The other standard on the market is OpenID by OpenID Foundation. You may get and use your OpenID in various services enlisted on the page. OpenID is more widely distributed standard than CardSpace, but… Microsoft is not willing to compete with it, but they has announced they are willing to integrate! Furthermore myOpenID mentioned above seems to be already the service integrating both possibilities CardSpace and OpenID.

If you are willing to use the standard you should visit the OpenID source page (OIDS), where you find all necessary downloads (documentation is not available though) and information about on-going projects. Generally there are two ranks of participants:
- OP : OpenID providers (the plan is to have at least one per country, but the coverage is currently quite poor) – it is counterpart of MS’s IP
- RP : Relying party (the users of OpenID basing on OIDS license – for example web site owners)
The whole idea seems to be a good, open campaign to organize chain of trust between the identity providers and web sites owners, to avoid the uncommon phishing attacks like the one described in here. There is quite intriguing business model presented by the organization:

Fig4 – How OIDS works

Future
It seems like OpenID is a step forward CardSpace as it starts to build already its net of providers, users and partners. If you write in Google “How to become Identity Provider”, among 10 first findings, there is no information about CardSpace or any other Microsoft service, but there are two linkages (VeriSign and Public OpenID providers) linked directly with OpenID enterprise. If there is not some visible promotion of Information Card providers the whole idea may end in a similar way to Microsoft Passport – it will be used mainly on Microsoft pages J

The other problem is that there is not too many visible places where you can get the Managed Information Card. Even the one, which you can find on the web does not work often, like the one. Furthermore even basic Personal Information Card does not work in some web services, like SignOn. Funny thing is that it works for some people or it worked in the past. Of course, the error might be (and probably is) in web site, but annoyed non-techie people may blame CardSpace.

The interesting option though for Microsoft is to provide “presentation layer” and use the OpenID as a “transport layer”. As there is cooperation planned for both and myOpenID already goes in that direction it is quite possible that future Windows (Vista) user my authenticate with CardSpace to get widely distributed and popular OpenID in the network.

piątek, 23 listopada 2007

Badmail coding adventure

Mood: Saturday party - heeeeaaaar I come !!!!!
Soundtrack: "Monkey dance" remixed (no comments)

Introduction
In one of the projects, we had a problem to ensure that the sent email, was received. As it is easy to guess there is number of reasons, why the thing may fail and one of the most obvious reasons is the wrong email address. The standard SMTP mechanisms does not provide any solution as it is assumed that is more Mail Server problem, than a protocol itself. I use the standard IIS SMTP preconfigured in well known way, but there is still couple of ways, how you can deal with a problem.

First of all you need some simple application to send the emails with use of SMTP service. To save the time you may use one of the application available on the web again, like the one. Then there are following options…

Custom headers
There is quite cool service on the web describing how the System.Net.Mail library works and at one of the paragraphs describes that

By adding custom headers, we can tag email messages with information that may not be visible to the end user, yet still might be important to us. This is especially useful in tracking bounced emails.

Well this is pretty close to what one need. Adding invisible header and reading it is not a problem with one exception… Where will we get the information that the email is bounced?

Badmail
This is a specific folder configurable in Messages>Badmail directory (by default C:\Inetpub\mailroot\Badmail), where all non-send emails are delivered after the predefined number of unsuccessful tries. You may easily test it sending the email to non existing client and reseting IIS service – the email will appear in Badmail directory immediately (usually you must to have a while).

Now, each bounced email will generate 3 files in fact:
- BAD mail which is the bounced email packed with leading message and delivery status
- BDP, which contains binary information related to the file
- BDR, which contains information about why the file was not deliverable

In fact what we need is the BAD file analysis – the plan minimum is just look for some unique name of custom header and check its content defined during sending the email as the error message ready to display.

Well… that is not elegant solution, so it is much better to parse fully the BAD file.

MIME reader
Lastly, on CodeProject there was released .Net POP3 MIME Client library, which is ready to use library – plug & play. You just need to make one small change in MIMEReader library and change one of the conditions in ParseBody:

//Parse a new child mime part.
[…]
else if (string.Equals(_entity.ContentType.MediaType, MediaTypes.MessageRfc822, StringComparison.InvariantCultureIgnoreCase)
&& ((_entity.ContentDisposition == null)
string.Equals(_entity.ContentDisposition.DispositionType, DispositionTypeNames.Attachment, StringComparison.InvariantCultureIgnoreCase)))
{
[…]


Now you are ready to create your own function to parse all bad files from bad folder…

public static ArrayList GetShortListOfBadmails(string badMailFolder)
{
ArrayList result = new ArrayList();

//Get all BAD files in Bad emails folder
System.IO.DirectoryInfo dir = new System.IO.DirectoryInfo(badMailFolder);
foreach (System.IO.FileInfo f in dir.GetFiles("*.bad"))
{
System.IO.StreamReader s = f.OpenText();
string[] strArr = StreamToArray(s);

MimeReader mimeReader = new Net.Mime.MimeReader(strArr);
MimeEntity mimeEntity = mimeReader.CreateMimeEntity();
//MailMessageEx mimeEmial = mimeEntity.ToMailMessageEx();
/*
* Children[0] is the leading message (text)
* Children[1] is the "Reporting-MTA: dns;akoszlajda-lap.pwpw.pl" message (delivery status)
* Children[2] is the email which was supposed to be sent (rfc822)
*/
int i = 0;
bool rfc822Found = false;
for (i = 0; i < rfc822found =" string.Equals(mimeEntity.Children[i].ContentType.MediaType,"> 0))
{
MimeEntity nonSentEmail = mimeEntity.Children[i].Children[0];
try
{
string strErrorDate = mimeEntity.Headers.GetValues("Date")[0];
string strTo = nonSentEmail.Headers.GetValues("To")[0];
string strEmailDate = nonSentEmail.Headers.GetValues("Date")[0];
string err = strErrorDate + " - Mail wysłany do " + strTo + " " + strEmailDate + " nie został dostarczony";
result.Add(err);
}
catch
{
//if there is lack of the information do not do anything
}
}
s.Close();
}

return result;
}


Other option
Optionally the problem with bounced emails may be send to specified email address and you may parse the emails there using the .Net POP3 MIME Client library, but you must somehow filter that email is the report about bounced email. That may be a problem.

Conclusion
The solution is very narrow and it is rather the proof of concept. There are is a number of other reasons why the email is not delivered (like the target email box is full) and the solution was not tested with Microsoft Exchange, but there should not be a big difference as the system has also Badmail folder mechanism.

On request, the zipped solution may be provided.
PS.
I know that there are some significant obstacles with Badmail installation on Windows 2003 using standard SMTP! I am not quite sure if it is impossible or there is some specific workaround.

wtorek, 16 października 2007

5 ways to produce valuable software

Mood: Regular
Music: This time instead the music - watch this!

Answering a question posted on the LinkedIn

If you truely plan to create the valuable software you must remember that your targets are...

1. Focus on the solution - fix the problem!
Joel Spolsky said once that, if you plan to make software it must fix a REAL problem - rule number. The best possible, the coolest solution will be bad, if it is just a toy. If you truly make the others life easier and customer surveys are saying that it means your product is good.

2. Focus at users - be stable, intuitive and user friendly!
The solution should behave as we expect it to behave. Whatever you say the MS Office package is quite a good sample in here (definitely 2k and 2k3). As my old IT teacher said… if your mother can install it, without a question and without the manual plus she knows how to use it after some time without cross wording – it means you are there.

3. Focus on the customer - has good value/price factor!
Of course, the best solutions are free solutions, but I guess we speak about the commercial packages. There are some articles, which are saying that there is a rule: “the barrier between 0$ and 10$ is bigger, than the one between 10$ and 100$”. That is a reason why it is sometimes better to provide the complex solution, which comparing to the other, similar things provides simply more USEFUL functionalities (1) for smaller money. Notice that most of users, use just a couple of small tools and most of them are free.
Wise price list may be a key to success. I love solutions like ZoneAlarm or AVG, which are free for private use and cheap for business.

4. Focus on fun - be cool and fresh!
LinkedIn the best sample for me – it is so easy to see if the company worked with some professional graphic / GUI expert or not J

5. Focus on community - do the they space for sharing knowledge between users, listen to them and support them
It applies also to non-web solutions; SAP has a very strong users communities. Microsoft focus less on the centralized MSDN right now and more on the communities, which supports people answering the direct questions. If the product vendor takes care about this community, it grows and helps to improve the product – that is a part of the Web 2.0 movement.
If your product is big enough, it should provide some place for plug-ins and platform for external developers to write extensions (AppExchange from SalesForce is good example).

... and this is it. So simple and so hard in the daily life, where you need work out the common vision of your product; it is so easy then to loose some or all of these focuses :)

poniedziałek, 20 sierpnia 2007

Technical Documentation Generation

When you create the big software solution, you usually need the technical documentation. Since some time there exists the automated documentation builders like NDoc. As I had some previous experiences in VS.Net 2003 (.Net 1.1), I was quite sure it will be nice and quick task to do the thing. How big my surprise was, that new technologies does not always make people live easier! :)

Basic NDoc is no longer supported for VS.Net 2005! The last NDoc version 1.3.1 does not work for .Net 2.0. There is some way to change the configuration file, so .Net 2.0 solutions are supported, but this is rather workaround than a complex solution.

I decide to look for some other new “mainstream” solution and I have installed the Sandcastle, especially that there is quite a number of fellows talking about the solution – among others Tomasz Kopacz, who is local, polish tech guru. The main page to start exploring the whole topic is not so popular on Google search. Sandcastle has surprised me because it does not provide you the GUI within the main package and you must download some additional tool for it like the Sandcastle Help File Builder. Unfortunately there is the other back fire and you must have generated xml file with comments (XmlCommentsPath), which was not necessary in old good NDoc! There is some nice project, explaining the basics, however it supposed to be some simple task!

As expected, I am not the only one who has this type of problem and there is number of places, where this discussion repeats – Joel on Software is one of them. I started to look for some other solutions, but most of them like DocumentX or Doc-O-Matic are not for free. Doxygen looks quite odd. And then the whole loop closed…

NDoc 2.0 Alpha, published by Kyno Sarges, who truly did the tremendous job seems to be so far the best option for old NDoc fellows. It is quite sad that NDoc home page is DEAD and you must to look for the remedy in such a places, but it works. It beats Sandcastle definitely and furthermore it provides you the comments’ xml file as well, but the difference is IT IS OPTIONAL! Unfortunately it is no longer possible to consume the whole solution file and have the documentation ready in a minute, but… at least you can do something quickly, without the necessity of writing custom files per project.

There is a good document, which described most of common problems with NDoc2 Alpha. The major problem is though that NDoc2 is no longer supported! Dead body again :) There is also the other drawback - unlike in NDoc 1.3.1 you must generate the xml comment file for each project separately (!) and add it to compilation in order to get summary and parameters tags consumed. How to make a trick is described in details in the article and the other one. The positive thing is that you can easily switch on one the parameters so the final chm document shows you in read which summary or parameter is missing the content of the comment and in such a way you can easily find the places in code to fill in.

So this is how the whole story finishes – there is one thing, which I sure of. Creating technical documentation for software solutions takes now much more time and effort than it did before! :)
ps.
My programmers finally have chosen the Sandcastle as it is the only one supported by the Microsoft and they have convinced me. I am just after the second review of the generated technical documentation and I love the options showing on red what is missing ;)

czwartek, 28 czerwca 2007

„Evaluating Software Architectures: Methods and Case Studiem” book review

Mood: In hurry

Introduction
The book describes three methodologies of reviewing system architectures:
ATAM – Architecture Tradeoff Analysis Method
SAAM – Software Architecture Analysis Method
ARID – Active Reviews for Intermediate Designs
It concentrates mainly on the first method and afterwards it describes the following two much more briefly (especially ARID) basing on previous material.

Content of book
The biggest advantage of the book is a very practical approach to the very abstract topic. It is a direct guide for a novice person showing step-by-step how to:

  • prepare for evaluation
  • make evaluation
  • collect feedback from architecture review process

It includes the detail break down of the methods into phases and stages including the suggestions about audience, meeting agenda, topics to cover and threads including the ways of managing them. The authors of the book show a lot of real samples from particular stages of the process and provide number of very useful tips.

Mentioned methodologies do not cover the 360-degree review of the software architecture in particular project, but they concentrate on defining the main bottlenecks of the projects and following deliverables:

  • Creating the usage attributes tree [drawing 3.4, page 72] with assigning priorities to them - ATAM
  • Defining the most critical, common usage scenarios - ATAM, SAAM, ARID
    Creating the real risks list - ATAM
  • Defining quality attributes [drawings 5.1-5.3, page 132; table 6.3, page 190] and quality targets - ATAM
  • Synchronizing the knowledge about project and project vision among key persons in project - ATAM, SAAM, ARID
  • Reviewing available project documentation and architecture - ATAM, SAAM, ARID

Chapter 9 and Table 9.1 [page 274] contains the detail comparison of all three methodologies.

ATAM
ATAM methodology is the precisely structured (into phases and stages) extension of SAAM method and it is the best fit for:

  • Review of existing project before creating its new version or reuse in other project
  • Formal review by external authority of started or existing software project (for example before taking some strategic decisions)
  • Taking over the on-going or existing project from one team to the other
  • ATAM focuses on big and medium software projects. Tables 2.1 and 2.2 [page 61] show the approximate cost of evaluation process (medium – 36 man days, small – 15 man days).

SAAM
SAAM methodology [Drawing 7.1, page 234] is more flexible approach to topics covered by ATAM. It is also applicable to big and medium projects, which are during realization or before creating the new version. The main difference between SAAM and ATAM are:
SAAM is focused mainly on functionality and extensibility of the project (ATAM covers wider set of attributes - also security, scalability and reliability)
Slightly smaller cost of the evaluation (minimum 2 days comparing to 3 days for ATAM)
More iterative way of creating the deliverables – agenda of meetings is the same each day and cover topics from all the stages, unlike ATAM, where each stage has different agenda

ARID
ARID is the only methodology mentioned, which is applicable during (and not after) the analysis phase. Like SAAM it is based on iterative approach, where agenda (for phase 2) is the same through the whole time of the evaluation process. Unfortunately, this methodology described in a very brief way, without a sample.

The cost of using this methodology is similar to SAAM (minimum 2 mandays).

ADR
The common point of described methodologies is the active approach to the review process (Active Design Reviews, ADR) and bringing back the real effects from the process. The main reason of doing it is to guide the evaluation process so it does not provide just junky type of reports “everything-goes-fine;don’t-worry”. The genesis for this approach is the article “Active Design Reviews: Principles and Practises” written by David Parnas and David Weiss described in details on page 261. It says among others about proper way of putting the questions (non-rhetorical, but focused on descriptive answer), what brings back much better answers and results from review process.

Other methodologies
The book mentions also the number of the other evaluations methodologies (in Chapter 9) like:

  • Quality Attribute Workshop (page 286)
  • Software Performance Engineering (SPE)
  • Survivable Network Analysis (SNA)
    Etc.

They are not described in details and book is just a good reference point for further readings about them.

Source of Document Templates
Book provides also number of good template for number of documents like usage scenarios, usage scenario analysis [table 6.4, page 196], architecture approach [drawing 3.6, page 81], quality attributes, architecture drawings etc.

Opinion
It seems like authors of the book WANTED to create the full review of all “architecture review” approaches and they started it, with the full and a very nice ATAM overview, but afterwards it seems like they did not have enough steam to do the same for the other methodologies doing shorter and shorter descriptions in following chapters. At the end of the road, instead of a nice encyclopedia type of the book, it is in fact the book about ATAM and similar methodologies. It concentrates just on a middle and big projects, which already has started and does not give a lot of support for analysis phase (except short description ARID), where the most of this type of work SHOULD be done. Unfortunately, it seems like the reason of such approach is quite simple – the authors are selling their service and support in ATAM area.

There is number of references to other methodologies, but there is a visible lack of connection to software designing approaches like Agile methodologies, RUP, MSF or project management methodologies like Prince2 or PMBoK. Such two chapters would show much better where these processes could be applied during software development process.

Despite of the criticism, the book is definitely worth of reading as it is VERY practical and it shows a unique point of view for project managers, software architects, stakeholders and other key players in software development process. It provides a VERY good starting point for preparing the evaluation process, giving a number of nice tips; among others it says how to convince the stakeholders that the whole evaluation process is necessary and how to build the review architects team.

Final score: 4 (good)/6

Annotation
References to pages may be slightly different in English version of the book as the article bases on polish translation of the book made by Helion.

 
web metrics