Re: ImageJ, ImageJ1, and ImageJ2 entries in omics tools
On Fri, 21 Dec 2018 at 09:46, Fabien Pichon <fabien.pichon@omictools.com> wrote:
>
> Hi David,
>
>
> Concerning the link to https://imagej.net/ImageJ, what I wanted to
> show you is not about flavours, but you can notice in the first row
> of the table (ImageJ2), third column, that they explicitely wrote
> that developers of ImageJ2 are "ImageJ developers". So, in our mind,
> this is the same lab that developped both tools.
>
It's not the same people. If ImageJ1 and ImageJ2 are both ImageJ,
then both groups are ImageJ developers.
> The confusion, here, seems to araise from the mix of ImageJ and
> ImageJ1, which are not the same it seems...
>
> So, if I understood well:
>
> - ImageJ has been developped by developpers in 1997 inspired by NIH image,
> - ImageJ1 is ImageJ, but maintained by another (unique) developper,
> - ImageJA is ImageJ1 corrected to be Git-compatible (maintained by both ImageJ and ImageJ1 developpers),
> - ImageJ2 is developped by the same developpers as ImageJ (not ImageJ1), based on ImageJA.
>
> Whow ! That's complicated !
No, that is wrong. It's more complicated.
ImageJ was developed by a single person (with contributions from
community). This is the same person that developed NIH image. He
continues to develop it. Let's call this ImageJ1.
Another group started to develop an improved version of ImageJ. It's
also ImageJ and is ImageJ2.
ImageJ1 is not a very nice library to be used as part of another
program. The ImageJ2 developers have created ImageJA which makes it
easier to be used in other java programs with maven. ImageJ2
developers make use of ImageJA instead of ImageJ1.
Now, there are many plugins for ImageJ. My experience is that most
users don't realise they are using plugins if they come installed and
each distribution of this many "flavours" have different sets of
plugins.
ImageJ1 is released in both source and binary form. The binary form
comes with many plugins. The source release does not include the
source for the plugins. Debian and Ubuntu build ImageJ from source
and so those "versions" are less capable (because the typical user
will not realise this is a plugin, even if they appear in the plugin
submenu).
> So, do we actually need 4 entries ? ;)
>
> As I said, we are not experts in all fields so if you are an expert
> and you say that 2 entries would be better, we will discuss
> internally to add a new one. No problem. Today, the best reason I
> see to add a new entry, definitively, is the license. Because this
> is particularly relevant for the tool use. But if you say that 2
> entries would be useless, we keep 1...
I guess it depends what's the purpose of the entry. For me its use is
not clear so I can't comment anymore whether you should have a
separate entry.
If license if a major thing --- it is for me --- then note that
ImageJ1 binary releases have a less clear license. The source in the
source releases is in the public domain but the binary releases
includes plugins whose sources are scattered online and whose license
would have to be checked. If people are using Omics to audit for
license, then this needs to be clear because I don't know anyone else
that actually uses the ImageJ1 source releases.
>
> I also note that we already have Fiji as a different entry, which is
> just ImageJ.... with some plugins already pre-installed.
Fiji is the main distribution of ImageJ and the one I recommend to my
users. Again, if you are using omics to audit license, you probably
should check the license of its plugins. It used to include some
non-free plugins (without a fee for non-profit).
Best,
David
Reply to: