What happens when I drop a file on archive in Nautilus?

Question

I had this file structure:

2000/
  001.txt
  002.txt
  003.txt
  ...
2001/
  ...
2002/
  ...
...
2011/

I guess file structure doesn't matter, but anyway...
So folders are representing couple of years with some text files inside.

Now I compress all, with Nautilus compress context menu (archive manager) in tar.gz format. I choose this format as although it requires more space then bz2, it's lot faster when I need to query data inside.

After this I create helper text file with references about archive contents (170 KB in size) and wanted to add it to archive, so I just drag & drop this file onto created archive in Nautilus.

Two bad things happen:

1. Droping file onto archive extracts whole archive in temp folder then re-compresses all files again! This is just awful for large archives, and I hope to find better way in Terminal later

2. Initial archive was ~ 96 MB in size, and now just by adding this 170 KB text file, archive become 140 MB! I checked inside it's same content + this tiny text file

Why is this?

---

I just created archive with contents from two subfolders, than drag&drop file with size of 100 bytes:

$ gzip -l tmp-original.tar.gz
         compressed        uncompressed  ratio uncompressed_name
           13297512            69908480  81.0% tmp-original.tar
$ gzip -l tmp-updated.tar.gz 
         compressed        uncompressed  ratio uncompressed_name
           19231941            69908480  72.5% tmp-updated.tar

Answer 1

As a note on point one, as a far as I can tell, the tar terminal command cannot append files to a compressed archive (ex. your .gz archive), and hence needs first to uncompress the archive, then add the appended file, and recompress again. A workaround could be mounting your compressed archive using archivemount (instructions here), which allows you to mount your compressed archive, append files, and unmount it. Whether this is speedier than uncompressing, appending and recompressing, I cannot say.