CMSMS and new versions

[Sy]

I've been using CMSMS for under a month now and in that time they're have been 4 releases...some might find this off-putting, updating it this frequently gives the impression that it isn't very stable.

I know that it is a good product and an very pleased with it, I would like to suggest that you consider controlled and staged release's of the product.

The last I heard the next version was going to be 1.2, then out of the blue comes 1.1.4, quickly followed by 1.1.4.1, and it was exactly the same with 1.1.3, followed by 1.1.3.1

You need to get hold of the quality control and make sure its well and truely tested before releasing...when you announce a release to the world, consider that many many people could render they're sites open to exploits or break down all together by implementing a recommended but rushed out release.

[calguy1000]

I nominate Sy to the quality control team. 

Then our weekends will be easier, knowing he'll test every upcoming release, and all the permutations and combinations.  He'll make sure that nothing 'accidentally' happens when we spend 3 hours cutting a release on a Thanksgiving sunday and one character error happens that messes it all up.

The other solution of course is to pay me or Ted a full time salary, and then we can have the time to do all of those tests ourselves.

[Sy]

I am a full time software developer, I know the problems faced with large projects and the knock on effects encountered by issuing software without fully testing it.

I have done it myself two often...its something I try to avoid.

I don't think anyone is holding a gun to your heads pressurising you to release versions early, and don't get me wrong I am not being anti....what I am asking for is some confidence in that the version I have just updated my site to is stable and free from exploits to the best of your knowledge.

I know you can't give any guarantees, but you should be able to capture the kind of bugs that you have found within 24 hours of releasing an update.

[calguy1000]

Ted, _SjG, myself, and other members of the development team are also full time developers.  We don't enjoy sending out new releases all the time (it takes anywhere from an hour or more just to cut the release, upload it, edit the downloads page, and announce it).  It's not an easy process.

We focus on security fixes, and make every effort to get security fixes out as soon as possible (hence 1.1.3 and 1.1.4).  In the both cases we received an email from valued contributors that let us know of some new, and longer standing security issues that absolutely needed to be fixed.

However, tiny mistakes slipped in in both cases, and in less than an hour in both cases we'd rectified the problem (1.1.4 .1 took a little longer to upload), however it required 1.1.3.1 and 1.1.4.1 respectively.  We have a good user community here, they spot these issues in less than an hour most times.

1.2 was announced a month ago, and has been in beta for over a week.  We couldn't rush the release of 1.2 so we had to release a 1.1.4 it was as simple as that.

I've seen commercial packages that were in 'beta' for five to six months, with three or four beta packages all of a sudden crater when the final release went out, requiring emergency 'panic' fixes.  There are no guarantees that even a well tested, well examined, package with an ultra-strict development policy won't run into these problems (they're just minimized). 

And we've been on a bad stretch lately. I'll get better at releasing the software

[Signex]

The Devs try their best that as soon as security breaches are found to issue an update asap. Most times security releases are tiny fixes so testing is not always needed and the mistakes with 1.1.3 to 1.1.3.1  and 1.1.4 to 1.1.4.1 are the only 2 times this happened at least as far as I can remember, I`ve been using cmsms since 0.10 or something. I prefer this rather then waiting another 2 weeks for 1.2 because else they will ruin the release circle.

[Sy]

It is a difficult balance to strike...I work in the offshore oil and gas industry, when something goes wrong it has to be fixed as soon as possible because it can cost upto several hundreds of thousands for one days down time.

Sometimes we have no choice but to hack something together as quick as possible and get it out there.

We have also ended up fighting more fires this way as when one thing goes wrong, everything else decides to go wrong at that precise moment.

Please don't take my post as having a pop at anyone, as its the last thing I wanted to do...what I was trying to suggest is that releasing new versions frequently does not create an impression of stability...perhaps these newer smaller releases should be called HotFixes instead (Sounds like something you get from Microsoft), a hot fix is different form a new version...which is what the Announcement on the main page makes you think.

You can have any number of hot fixes to address any number of issues.

[calguy1000]

That point I'll agree with, maybe we should come up with a release naming standard of some sort).

And... I worked in the oil & gas industry (offshort & onshore) for like 12 years or more).  Working with companies like paradigm, and photon, and cogniseis..... got any openings

[Sy]

lol, we are the same age, I'm in the UK though, currently working for a French Company, Perenco UK Ltd.

I've been working for the offshore industry in various areas all my working life...started out in the survey industry writing, pre and post processing software...then started writing communications drivers for GPS, Gyro's, Altimeters, various sensors, then got into control systems and wrote some ROV software, moved onto sub-sea writing sub-sea networks using Lonworks (echelon).

Now writing software for Hydrocarbon Gas Allocation and Metering...also developing AJAX Intranet for Perenco UK.

In my spare time I code, developing sites for friends etc.

[calguy1000]

France, I could do France

My experience is mostly on the upstream side.... exploration and modeling stuff.  Basemaps, synthetic seismograms, rock physics and all that stuff (though I am no scientist, they write the pseudo code, I work with them till it works, is commercially viable, etc, etc, etc).

[Pierre M.]

Hello,

....what I am asking for is some confidence in that the version I have just updated my site to is stable and free from exploits to the best of your knowledge.

Very legitimate. Have a look at RonnyK's Major.minor.bugfixlevel numbering scheme and release policy idea.

Pierre M.