XYplorer Beta Club

I am a hardcore user of color filters. But sometimes color is not the most useful way to rapidly identify files in a list.

I was thinking about using the text size instead like in a "search cloud" that can be found on some web pages (see http://johnbokma.com/perl/google-search-cloud.html).

The rules would be the following:
-define max font size for the most relevant
-define min font size for the least relevant (with a threshold to be defined here. this could be a date for example: for files older than this date, the font size should be equal to the min font size)

Example:
Very_relevant1.txt less_relevant.txt not_so_relevant1.txt Very_relevant2.txt not_so_relevant2.txt

On top of that, increasing transparency could be used for less relevant files so they finally merge with the background.

Example:
Very_relevant1.txt less_relevant.txt not_so_relevant1.txt Very_relevant2.txt not_so_relevant2.txt


Cheers!

Funny: yesterday I thought about adding a recently (or frequently!) used files list organized as a search cloud! Or a new Catalog format, which is not just a list but a 2-dimensional field where you can freely position your items in 2D-space.

But, please give concrete example. I don't fully understand the term "relevant" here. What kind of files are you searching, by which criteria?

If I am interested in the most recently created files, the relevancy would be the age of the files.
The youngest, the most relevant. The oldest, the less relevant.
I chose the age in this example but it could be the latest modif time, the size etc...

I hope it makes it clear

something like a tag cloud is a good idea. what window exp calls it frequently used files/folders.

noutters wrote:If I am interested in the most recently created files, the relevancy would be the age of the files.
The youngest, the most relevant. The oldest, the less relevant.
I chose the age in this example but it could be the latest modif time, the size etc...

I hope it makes it clear

But this is just 1 dimesion and can be done by sorting. I thought tag clouds are meant to represent more complex relationsships, especially if you mix various attributes like size, color, brightness, distance...

So, what would be the advantage over simple sorting in your example?

In my opinion,

say i have 3 drives, c, d and e:

c is for programs, d for work and e for fun.

talking about clouds, at work my d drive cloud will be bigger than at home where e would be bigger. also at home, certain folder such as music might have bigger cloud than say photos.

more the number of clicks on a drive bigger the cloud, i think for one needs a frequently used tasks normally for one or two days (somethign like a working history) but in general the emost popular folders should also be highlighted.

In short, i can see myself using both frequently used folders and most popular folders.

eg:
1. i use xybeta folder 3 times a week and over time it becomes popular. its not recently used but is popular.
2. when i am on a project i need my latest used files to be accesible which need not be popular or useful in a week or so.

update: maybe in catalog you can have frequent and popular categories, which will populate automatically based on what user wants. frequent is self-understood, but you can provide for example an option to use only d and e drive in frequent. on other hand in popular, most clicked one stays on top.

serendipity wrote:update: maybe in catalog you can have frequent and popular categories, which will populate automatically based on what user wants.

I call it frequent (stats over all time) and recent (stats over say last week). That would be 2 quite useful automatic clouds. Something to meditate about...

Well, I don't think I really like clouds myself, so....

But what I'd like would be the ability to use, for color filters:
- more "colors", ie. text but also background colors (with some blending/mixing for the odd row/zebra stripping).
- more "filters", like criteria based on dates or size

Ideally, we could choose to have some of those to apply only on a defined column, eg I could say *.rar are on blue, and >700MB have their size column red. That way not only does my RAR files are in blue, but I can also spot the big ones cause their size column stands out in red.
And then also, the Created column is in green if the file was created in less than 24h, but is purple is it was last month or before !

admin wrote:

serendipity wrote:update: maybe in catalog you can have frequent and popular categories, which will populate automatically based on what user wants.

I call it frequent (stats over all time) and recent (stats over say last week). That would be 2 quite useful automatic clouds. Something to meditate about...

Definitely a good idea here! That would allow one to have a catalog group that keeps itself maintained based on my actions, as opposed to having to do it manually.

I'd suggest three user settable parameters as minimum required:
1) Time period for each, as user may want frequent to be over last month and recent over last 48 hrs, for example.
2) Max number of entries in each, so they don't get out of control.
3) Number of usages before it gets added.

A number of things to consider also:
1) Block adding to these if already in catalog elsewhere or not?
2) Cleanup of obsolete entries when some are deleted/renamed

I also agree with jacky's ideas on expanding filtering abilities, but the tag cloud is one thing that I'd never have a need for, as I find them very distracting.

jacky wrote:But what I'd like would be the ability to use, for color filters:
- more "colors", ie. text but also background colors (with some blending/mixing for the odd row/zebra stripping).
- more "filters", like criteria based on dates or size

Ideally, we could choose to have some of those to apply only on a defined column, eg I could say *.rar are on blue, and >700MB have their size column red. That way not only does my RAR files are in blue, but I can also spot the big ones cause their size column stands out in red.
And then also, the Created column is in green if the file was created in less than 24h, but is purple is it was last month or before !

Yep, not bad. Make a mental note and bring it up again for say ver 7.0 ...

j_c_hallgren wrote:

admin wrote:

serendipity wrote:update: maybe in catalog you can have frequent and popular categories, which will populate automatically based on what user wants.

I call it frequent (stats over all time) and recent (stats over say last week). That would be 2 quite useful automatic clouds. Something to meditate about...

Definitely a good idea here! That would allow one to have a catalog group that keeps itself maintained based on my actions, as opposed to having to do it manually.

I'd suggest three user settable parameters as minimum required:
1) Time period for each, as user may want frequent to be over last month and recent over last 48 hrs, for example.
2) Max number of entries in each, so they don't get out of control.
3) Number of usages before it gets added.

A number of things to consider also:
1) Block adding to these if already in catalog elsewhere or not?
2) Cleanup of obsolete entries when some are deleted/renamed

Yep, all this makes sense. Although I'd like to start it as easy as possible before getting lost in configuration jungle... ... but it will be months until I have time for this, so there's plenty of hours for meditation...

Some very interesting work for representing files on a disk is the concept of treemaps. See this link for info: http://www.cs.umd.edu/hcil/treemap-history/index.shtml

I have been looking around at different applications that generate treemaps of drives or directories. IMHO, the best result is achieved in Windows by SpaceMonger. It can be used almost like a file browser by zooming in and out the directories.

This is a radical new way of browsing since browsing is made over the whole disk or an area representing several directories.

Think Google Earth for your hard drives.

It misses however all the major great features of XYplorer.
In XYplorer, the size of boxes could represent the age, or the size, or the most recently used files etc...

Here is an illustration:
http://www.sixty-five.cc/sm/img/treemap.gif

What do you think?

noutters wrote:Here is an illustration:
...
What do you think?

I think the image is blocked.

PS: but you can see it via direct browsing.

Well. It looks like an (interesting) experiment in progress. I think human brains are very good in memorizing paths through landscapes (topology, landmarks...) and in memorizing faces, voices and smells. Why not go all the way, and make it even more like goo earth!? I think in future, the harddisk will be represented like a virtual room or house or garden or landscape, totally realistic (think second life), plus sound and smell.
Maybe a bit far out for now. There are still some old schoolier things on my road map...

I've tried a appl that creates a treemap of my drive and my conclusion:

It may work for some folks but NOT for this old guy who worked on IBM mainframes for 25+- yrs!

The concept of a auto-updating catalog group based on usage would, IMO, be much more of a selling point to new users...I don't often use competing products so have no idea if that is a feature others have or if it could be another unique XY thing!

And yes, agreed that we have plenty of time to develop it, but just wanted to start thowing out a few ideas about it now...here's another one: Have the ability to create multiple auto-update groups that are restricted to a particular parent, so one could, let's say, have one group for C: and another for D:Music instead of one that has everything in it...

admin wrote:Yep, not bad. Make a mental note and bring it up again for say ver 7.0 ...

not bad.. Not Bad.., freaking GENIUS you mean!!
Damn, v7.. hmm, maybe I'll try to trick you and try remind you around v6.8 instead

Me making a mental note is a sure way to forget about it , so I wrote it down on my little XY list intead (I had already done so in fact)

j_c_hallgren wrote:It may work for some folks but NOT for this old guy who worked on IBM mainframes for 25+- yrs!

I never worked on IBM mainframes, so it is not related I think as I don't like it either
I've never been a fan of clouds, and this doesn't really do it for me either.

Thinking about the use of such things, I would say a way to see, for example, what are the oldest bigger folders somewhere could be an interresting thing.
But instead of clouds or that "new design/GUI" thing, I would go for a rather different approach : the ability to have XY sort a List using more than one criteria.

And I don't mean, by dates, and then when dates are the same, by size. No I mean to take into account both and "mix them up". So it's not the oldest items that would be first listed, but the ones that are from the oldest and are quite large.
Each criteria would be taken into account as importantly as the other, or each could have its own user defined level of importance.

That way, you could have a List in which the first items wouldn't be the biggest ones, or the oldest ones, but the ones that are amongst the oldest and biggest ones at the same time.


Regarding the ideas of "special categories" for Catalog, it could be an interresting feature, also I'm not sure why the "recent" one would be that interresting, I mean wouldn't it be like some kind of History?

The over all "frequent" might be more interresting, altough to be honest I'm pretty sure I wouldn't use it or like it so much. Just because I know what my frequent location are, and I add them to my Catalog myself.
So I decide which are worth be there and which aren't, and I can fully customize the whole thing (colors, label, position...)

I would only see an interest into this as a "guide", to see if I forgot and hadn't realize some location could/should be added on Catalog, as I use it often. But that's all.

Of couse that's only me