Some unfair comments

Juan Conejero

PTeam Member
Sep 2, 2004
7,402
67
56
Valencia, Spain
pixinsight.com
Hi all,

We have recently been informed about the proliferation of a series of comments on various social networks and forums about the latest versions of PixInsight that we have published, and especially about the just-released 1.8.8-3 version.

The comments in question come to say that we do not properly test the new versions that we publish, and that we release them with the intention that our users test our software so that they do the job of detecting bugs and issues instead of us.

Not only are all of these comments completely wrong, but they denote an absolute lack of understanding about what PixInsight is and what this project has been representing for more than 15 years. PixInsight is professional software development. This is neither a hobby nor a game for us; it is our very hard daily work, where we invest a very significant part of our existence. If someone thinks that we really do that, we invite him or her to use other products in the market instead of PixInsight, since they will surely seem much better and much better developed.

Besides that, these comments bother me a lot, personally, because for more than a year we have intensified our work even more, striving very hard and giving our best to make PixInsight a stable and powerful platform with an extraordinary potential of future development on four operating systems. I am very proud of the work done recently, not just by me, but also by a bunch of great developers and fine persons, who are contributing their fantastic work for free, so this kind of comment is particularly painful. We test every new version we release in the most comprehensive and intense way possible with our available resources on all supported platforms. If we make mistakes or some problems go unnoticed, we apologize for that and you can be sure that it is not because we have not tried to do our best to solve or detect them.

Other comments imply that we should release new versions less frequently, in order to test them more, instead of causing problems to our users.

These comments denote lack of understanding of what modern software development actually is, particularly for a dynamic project like PixInsight, which is evolving quickly and constantly, and which provides powerful open-source development frameworks to incorporate numerous contributions of collaborators and third-party developers. The people who contribute their scripts and modules are undoubtedly our best asset, and deserve all our effort to provide them with the best tools we can offer. Once more, this is professional software development, so we know very well when a new version is necessary. We do not release a new version because we want to try out new things, but because we are completely sure that the new version will make PixInsight more useful, more stable and efficient, and more enjoyable to all of our users. We are not perfect so we make mistakes, and our mistakes usually cause problems, but those who know us know that we always try to do our best to solve them as quickly and efficiently as possible.

Thank you for your attention.
 

astroedo

Well-known member
Aug 16, 2012
171
2
51
Italy
www.radice.biz
Thank you Juan fot these clarifications.

I've re-published your post on our Facebook Group (I don't know if you have a facebook profile, but here's the link).

https://www.facebook.com/groups/959090060876536/permalink/2670939369691588/

Clear skies

Edoardo
 

bilgebay

Well-known member
Oct 6, 2011
86
2
59
T?rkiye - Marmaris and Istanbul
I don't know who these are but please be sure that the majority of the PI clan are feeling very lucky to have you, Vicent and all the other kind people who are devoting a big chunk of their lives to this project.

The naysayers will always be but don't let them pull your spirit down.

 

Ruediger

Member
Oct 3, 2019
14
3
Dear Juan,

Let me drop my 2 cents as being a SW professional myself: Do not get bothered about these people. These people do not know anything about running such a complex SW project. Very often they are frustrated because they cannot catch up with rapid development process. The simple answer is: If it is too fast for you, or you are afraid of (new) bugs stay with your old version and skip some newer versions. And if this is not an option for them: Please feel free to use any other SW. You are not tight to any tool. You can decide.
And everyone should be aware of: No releases, no improvements.
Personally I am happy to download and install any new version, since I know there is something great to test and use which makes my images better. Best example WBPP

Let me assure that the majority of users is not represented by this complaining minority.

Hang on!
 

astroedo

Well-known member
Aug 16, 2012
171
2
51
Italy
www.radice.biz
My 2 cents too:

Juan have you ever considered to create a moderately vast group of "competent" beta testers with a wide range of machine and configurations? (maybe it already exists and I don't know  :D ).
Obviously working on a voluntary basis, screaming "FOR THE GLORY!!!"  :D :D :D ;)  ;)  ;)

This will make development slower, it's clear, but the "risk" for common user would be mitigated.
 

Ruediger

Member
Oct 3, 2019
14
3
astroedo said:
My 2 cents too:

Juan have you ever considered to create a moderately vast group of "competent" beta testers with a wide range of machine and configurations? (maybe it already exists and I don't know  :D ).
Obviously working on a voluntary basis, screaming "FOR THE GLORY!!!"  :D :D :D ;)  ;)  ;)

This will make development slower, it's clear, but the "risk" for common user would be mitigated.
I think that would be difficult. You need a huge group of testers to cover a significant amount of combinations of hardware, OS and also users skills. Compared to to the amount of PI users in total this would require a big portion. And I doubt that a small group would seriously improve error level, since the obvious bug are crunched in internal testing.

But anyway: To release pre-release Betas for field test could calm the critical voices.
 

vega41

Member
Dec 5, 2014
6
0
I wanted to solve the problem drastically by formatting the hard drive and reinstalling win10 and pixinsight. All the problems I had were solved!
The PC was a bit clogged with various programs and procedures and I think win10 went haywire.
Thanks
 

Juan Conejero

PTeam Member
Sep 2, 2004
7,402
67
56
Valencia, Spain
pixinsight.com
Hi all,

Thank you so much for your support and encouragement!

I think that would be difficult. You need a huge group of testers to cover a significant amount of combinations of hardware, OS and also users skills. Compared to to the amount of PI users in total this would require a big portion. And I doubt that a small group would seriously improve error level, since the obvious bug are crunched in internal testing.
Yes, things are that difficult for a large and complex project like PixInsight.

Anyway, I'm going to think on your advice of releasing public betas, or release candidate versions. This would probably increase the involvement of our users in PixInsight as a dynamic development project. I'll think on how to implement this.

I wanted to solve the problem drastically by formatting the hard drive and reinstalling win10 and pixinsight. All the problems I had were solved!
Now that's the attitude! ;)
 

mschuster

PTeam Member
Nov 27, 2011
1,130
4
San Francisco
Juan Conejero said:
Anyway, I'm going to think on your advice of releasing public betas, or release candidate versions. This would probably increase the involvement of our users in PixInsight as a dynamic development project. I'll think on how to implement this.
Thanks Juan. If possible, please allow installation of candidate without removal of the prior working version. To allow for testing, yet still be able to easily switch to the prior version to get work done if necessary.
 

Juan Conejero

PTeam Member
Sep 2, 2004
7,402
67
56
Valencia, Spain
pixinsight.com
Hi Mike,

mschuster said:
Juan Conejero said:
Anyway, I'm going to think on your advice of releasing public betas, or release candidate versions. This would probably increase the involvement of our users in PixInsight as a dynamic development project. I'll think on how to implement this.
Thanks Juan. If possible, please allow installation of candidate without removal of the prior working version. To allow for testing, yet still be able to easily switch to the prior version to get work done if necessary.
Interesting. Ideally we should have the possibility to perform a side-by-side installation. This is not possible right now. The main problem is the way application settings are being generated and stored, but it is doable. Thank you for the idea.
 

Ruediger

Member
Oct 3, 2019
14
3
Juan Conejero said:
Hi Mike,

mschuster said:
Juan Conejero said:
Anyway, I'm going to think on your advice of releasing public betas, or release candidate versions. This would probably increase the involvement of our users in PixInsight as a dynamic development project. I'll think on how to implement this.
Thanks Juan. If possible, please allow installation of candidate without removal of the prior working version. To allow for testing, yet still be able to easily switch to the prior version to get work done if necessary.
Interesting. Ideally we should have the possibility to perform a side-by-side installation. This is not possible right now. The main problem is the way application settings are being generated and stored, but it is doable. Thank you for the idea.
But I think there is some crucial points to consider: When using a side by side installation (which I is a good idea in my point of view) most users want to use there existing settings. They do not want to re-configure everything. So you have to clone the settings during install process or at least to ask.

Also the users who want to update there installation in place instead of having a side-by-side installation have to remain the chance to migrate there settings.

I thinks this needs some brainstorming because there are some use cases to consider in order to avoid collateral damage - and again complaints.
 

astroedo

Well-known member
Aug 16, 2012
171
2
51
Italy
www.radice.biz
Ruediger said:
I thinks this needs some brainstorming because there are some use cases to consider in order to avoid collateral damage - and again complaints.
I'd use the rule: you're a beta tester? so NO COMPLAIN!!! 
Only feedback!

;) ;) ;)
 

Ruediger

Member
Oct 3, 2019
14
3
astroedo said:
Ruediger said:
I thinks this needs some brainstorming because there are some use cases to consider in order to avoid collateral damage - and again complaints.
I'd use the rule: you're a beta tester? so NO COMPLAIN!!! 
Only feedback!

;) ;) ;)
I doubt that this is the right approach. Do you think you will encourage a large field test group to participate with that attitude?
Moreover no one is willing to test, if you harm or endanger their productive environment. How many would set up a secondary PC only for doing beta testing?
That is something you can expect from paid professional tester, but not the majority of a product with comparable small user group.

I think the easier and safer you can parallely test new versions, the more are willing to do so and contribute - and that is the aim of the game.
 

wadeh237

Active member
Mar 5, 2014
41
1
Ruediger said:
astroedo said:
Ruediger said:
I thinks this needs some brainstorming because there are some use cases to consider in order to avoid collateral damage - and again complaints.
I'd use the rule: you're a beta tester? so NO COMPLAIN!!! 
Only feedback!

;) ;) ;)
I doubt that this is the right approach. Do you think you will encourage a large field test group to participate with that attitude?
Moreover no one is willing to test, if you harm or endanger their productive environment. How many would set up a secondary PC only for doing beta testing?
That is something you can expect from paid professional tester, but not the majority of a product with comparable small user group.

I think the easier and safer you can parallely test new versions, the more are willing to do so and contribute - and that is the aim of the game.
On the contrary, beta testers are...testers.

In order to be an effective beta tester, you need to put aside your own goals and actually test the product.  This means that you should be keeping accurate records of what you are doing.  When you come across a problem, you should be able to do some level of investigation yourself, so that you can create an actionable bug report for the development team.  Depending on the needs of the development team, you may be able to do your own work, with your own workflow.  But you should be prepared to set aside your own work to do various tasks as requested by the development team.

I am a professional software developer who's run a number of large scale beta tests.  Testers who fall into the above category get lots of our attention.  We try to discourage folks who are mainly interested in early access to the "latest and greatest".  We inevitably get them, and they are useful for "taking the temperature" of the product to see how much noise they generate.  But it is a lot of work to follow up on "complaints" without sufficient supporting documentation to investigate the issues.  With a small enough development team, it might actually hinder product development.

I've also participated on a very limited number of beta tests for astronomy software.  In one case, I put my own imaging goals on hold for almost a year so that I could focus on the software I was testing.  It was work on my part, but it made the end product better, which was what I wanted to help with.

As for PixInsight, I think that it's an impressive accomplishment.  I know what goes into making a product like this, and my hat is off to Juan and his team.  There is a huge amount of hidden work that goes into making something like this work in a cross platform environment, especially when one of the platforms is Windows.  While it's true that there have been some stability problems with Windows in the recent versions, it looks very much to me like Juan is making the right changes to resolve them (and to be fair, most of the the root problems aren't even in PixInsight itself, but in 3rd party libraries on which it depends).
 

Ruediger

Member
Oct 3, 2019
14
3
Hi Wadeh237,

I agree to some extent to your arguments. But your argumentation matches mainly to a huge tester group. When we do a field test, which means 10 to 20 percent of the users (equals to 15,000 to 30,000 participants) we follow your arguments too. But usually you isolate the development team from the feedback. This is the task of the quality management team, and not of the developers. They have to consolidate and validate the feedback and channel it to the dev team.

But this is not applicable here. Juan wants to encourage people to test, and that you can only achieve IMHO if you keep the hurdles low. That is my experience from many small projects with strictly limited resources. Otherwise you will get only some hard core users.

I think it is also required to distinguish between a beta test in a relatively small group, or a large field test with release candidates.

BTW it was only my personal opinion and experience. In the end Juan has to decide and try it out himself, so there is absolute no need to start a discussion. It was only to give some perspective.
We should focus on supporting Juan as good as we can in the way it helps Juan most - whatever way he goes. 😉
 

CCD1024

Well-known member
Sep 22, 2007
403
0
France - Paris
www.cielaustral.com
Don't worry Juan, we are a lot of PI users who really love Pixinsight.
these guys are "fast food" eaters. They want all features for free and their big problem is they take poor quality images and they think PI will give them a nice image from their poor computer with fully cracked software... no way.
Fore sure, while developing, you cannot be sure all bugs will be fully corrected when publishing release. But the main point is your reactive action when a bug is discovered : corrected in few hours or days.

I would like to congratulate you and your team for 1.8.8.x version with new process and algorithms and also stability (I am under OSX)


Philippe,
a Pixinsight user since v1.0 for first version (even before for Denoise ACDNR) then 1.2 or 1.3 for license version.
 

wadeh237

Active member
Mar 5, 2014
41
1
Ruediger said:
BTW it was only my personal opinion and experience. In the end Juan has to decide and try it out himself, so there is absolute no need to start a discussion. It was only to give some perspective.
We should focus on supporting Juan as good as we can in the way it helps Juan most - whatever way he goes. 😉
No worries.  My response was my personal opinion as well, to add a counter point for additional thought.

I think that we are mostly in agreement, especially in your last sentence above.