Dear Kay,

Many thanks for nice words! We appreciate feedback from all our users.

Indeed, we expect that the update system will result in a lower frequency of releases. Yet, full-scale releases will be needed. A good analogy here are the update vs upgrade procedures in operating systems. An upgrade is needed whenever we face the need for structural changes in the Suite, which cannot be easily accommodated by a mere update due to a considerable number of various dependencies to be followed.

At present, we still need full releases because:

1) currently we are doing a considerable work "under the hood" of the CCP4 Suite, which is not reflected in what a general user would see. Some elements of that include better organised source repositories, synced with updates, completely new build mechanism, which would allow one to update installations compiled from sources, new test system and some others. From time to time, this results in structural changes in the backyard of CCP4 suite, which cannot be accommodated by update mechanism

2) sometimes we have new contributions which change really a lot. For example, when new GUI comes to release, it will change about everything

3) sometimes, new versions of various components result in changes far beyond their own scope. Usually (but not always), this relates to new versions of phaser, cctbx, refmac and some other programs 

4) incremental updates make the Suite asynchronous on component versions. For example, recent phaser update resulted in duplication of cctbx, which was necessary in order to keep other cctbx-related components running. With time, this results in higher complexity and lower manageability of the suite.

I am sure that this list is far from complete, hence one more reason

5) unpredictable things happen, when updater cannot manage something.

I hope that these reflections give some insight into maters and you will be less disappointed by having to upgrade to 6.3.1 fairly soon :)


Best regards,

Eugene



On 20 Jan 2013, at 11:32, Kay Diederichs wrote:

> Hi Andrey, Eugene and others -
> 
> it is time to say that this new system is extremely useful, and so easy to use at the same time! Gone are the times when one had to worry about manually updating parts of CCP4. 
> I just wonder why there should ever be a need for a CCP4 6.4.0 release; a rolling update seems just adequate to me - but that's just a user's perspective, and probably this has been discussed somewhere although I'm not aware of it.
> 
> thanks for making this available, and the hard work to make it so perfect!
> 
> Kay
> 
> On Fri, 18 Jan 2013 19:04:37 +0000, Andrey Lebedev <[log in to unmask]> wrote:
> 
>> Dear CCP4 Users
>> 
>> A CCP4 update has just been released, consisting of the following changes.
>> 
>> All systems:
>> 
>> *   Ctruncate: Introduced DNA/RNA reference curve; uses flat prior on request, or when have tNCS or twinning
>> *   Othercell: Various bug fixes
>> *   PISA, QtPISA: Stability fixes and generation of Remark 350 in no-complex situations
>> *   PDBCur: Finish gracefully, rather than terminate, when atom selection is empty
>> *   PDB_Merge: Various bug fixes
>> *   Xia2: Correction to Ctruncate wrapper
>> 
>> Linux and Mac only: *)
>> 
>> *   Aimless: Various bug fixes
>> *   Pointless: Correction for XDS input
>> *   Scala: Corrected output for CC(1/2)
>> 
>> *)  Corresponding changes for Windows have been already included in update No 13
>> 
>> Windows only:
>> 
>> *   Phaser.ensembler: Corrected path to syminfo.lib
>> 
>> 
>> If you do not currently receive updates, consider re-installing your CCP4 setup using the latest binary packages, which now have CCP4 Update manager (ccp4um) integrated.
>> 
>> Note that auto-updates will work correctly only with CCP4 release 6.3.0, therefore upgrade if necessary. Please report any bugs to [log in to unmask]<mailto:[log in to unmask]>
>> 
>> 
>> Many thanks for using CCP4.
>> 
>> Andrey Lebedev
>> 
>> 
>> -- 
> Scanned by iCritical.


-- 
Scanned by iCritical.