We determined it once again, once of too undoubtedly. One of our clients put on all alone (because the representatives for such situation work well in the same firm as that where officiates our interlocutor) in a losing / losing situation with its integral filter as part of a plan developed for a fixed sum. What are facts? We could, without the same let him speak, name one to one all difficulties which he meets, the closed eyes, like a rosary which is shelled.

A Call of Offer was issued for the main integral filters of the place because according to the Direction of Purchases it is the only way to lead a computer plan (We spoke you in a previous ticket:    ). It is, according to the same specialists admitted for their computer competences, a means to control costs and delays, to share risk with the chosen firm and to have a means of pressure (because of course superb clauses of penalties of delay were envisaged in the contract). The integral filter X is chosen. For a big part, because he had a tempting proposal and could accept, without saying word, all principles of this contract of confidence . The decor is planted: an atmosphere of serenity, a reciprocal confidence, the involvement of the best profiles of the society X, wishes of users taken into account

You suspect it, this last stage is ironic.

The project manager is under high pressure, his firm appointed him 3 and a half trainees, architect's quarter because, after bitter negotiations, earned affair turns out to be less profitable than to be envisaged. And inside in the Soci t X, a badly negotiated plan leads to the allocation of unskilled developers (  You pay peanuts, You get monkeys  , it is well known). The purchasers of the client rub their hands, they led well their small boat. The project manager is forced to make work everybody more than of reason to meet unreasonable deadlines. Word "penalty" murmurs discreetly then rings out for the record in all meetings of plan, having thrown out as supreme threat. The legal services of both parts take out again the contract and lean with a lot of interest there. Everybody passes besides more time to see how to get the other part from defect than to try really to resolve problems together, in good faith. And the users in all this? They get worried. The application which they ordered is critical for exercise of their job. MOE explains them that they badly worked, that their specifications are too vague, that their uninterrupted requests of modifications are unacceptable. They would be even checked off of the finger as second representative for such chaos, after the integral filter, of course. Ils seraient m me point s du doigt comme deuxi me responsable d un tel chaos, apr s l int grateur, bien entendu.

This plan quickly becomes a bottomless puit. The client went too much far to go astern. The box of Pandore is opened: "endorsements" , word is finally dropped to the great satisfaction of all because, in fine, it is the only way to take out the high head of both quoting. But attention, the usage of such blows of penknife in the contract linking both parts must be made with a lot of parsimony of course, because he must just allow the integral filter to reduce  his deficit "a bit " on plan, where really it is not possible to get him from defect. This subterfuge allows him, hardly, to stay in a vegetative financial statement on plan (reduced losses or at best no profitability).

The paradigm of software development for a fixed sum is based on principles (not to say stocks) counterfeit, dishonest, utopian, outdated.

The firms which succeed as Google, Amazon and Microsoft, to name no that the most knowing, included it for a long time. Agile methods are a serious alternative. The principles of these methods are the following:

• Competence and motivation of teams
  The personnel put in the disposition of plans is the level of seniority necessary for the development of modular, progressive, competitive, researched, integrable and exploitable applications. This personnel gets involved on ciphering and realisation of application.
• Iterative development
  The development of plans is iterative. Effect tunnel is abolished. Delivered modules are immediately operational.
• A favourable reception of change
  The writing of complete detailed specifications of the months before its delivery turns out to be an Utopia. In this context, the acceptance of the change of priorities, of functionality made integral part of the process of agile development of Xebia.
• A close collaboration with the users
  The involvement strong in management and in users throughout plan is organised around permanent feedbacks orchestrated by systematic demonstrations which allow, in every iteration, to validate done work and to define the priorities of following stage.
• An attention perms in technical excellence
  Hyperspecialization Popular dance / J2EE allows, from the beginning of plan, to set up the best practices in force in industry and it at all levels of plan: Architecture was conceived by experienced Architects, policy of integration goes on, performance tests, management of errors, exploitabilit of applications, permanent technological choices,