bobgr's list of bugfix requests

User 870ab5b546

02-09-2016 19:18:50

New version; some minor feature requests added, one removed. Please, please address the requests that are in bold!


Major bugs:



  1. If I replace an atom in a 3D structure with a shortcut group, the z-coordinate of the original atom is not transferred to the shortcut group. (This bug has caused students grief on exams.)

  2. Hypervalent heavy atoms, for example, a nine-electron Br atom, should not trigger a valence error.

  3. There is a problem with cross-compatibility of rectangle sides/midpoints between MarvinSketch and MarvinJS. If I draw a rectangle in one editor, then draw a reaction arrow that starts at the midpoint of its bottom edge, then display the structure in the other editor, the arrow starts at the top instead.

  4. If in MarvinSketch, I draw a rectangle A, then I draw an arrow B that begins at the midpoint of the bottom edge of A, then I draw an arrow C that begins at the midpoint of arrow B, then, when I import the structure into Marvin JS, arrow C disappears. Evidently, MarvinJS doesn't understand the concept of one arrow beginning at the midpoint of another arrow, so it just ignores it.


Minor bugs:



  1. Fix placement of charges so they don't overlap with bonds. (Again, this bug has caused students grief on exams.)

  2. When I draw a coordinate bond from atom P with a lone pair to another atom M, and lone pairs are visible, either the lone pair of P should not be displayed (preferred, and the current behavior in MarvinSketch), or it should be displayed at the starting point of the coordinate bond. 

  3. Using an iPad, for toolbar buttons with submenus (bond, arrow, electron-flow arrow, and rectangle/ellipse), is very difficult to find just the right spot on the lower right corner of the toolbar button to press and hold so that the submenus will appear.

  4. Using an iPad, even when I can get the submenus to appear, I am unable to select any of the options. For example, I cannot select the bold bond or the two-electron electron-flow arrow. Dragging my finger over to the desired option has no effect, and the submenus simply disappear when I lift my finger.

  5. Currently, I am unable to draw a graphical arrow from one rectangle corner/midpoint to another rectangle corner/midpoint if they are within a certain distance. 

  6. I can change the arc of an electron-flow arrow, but the arc parameters are evidently not stored in the MRV, because if I export the structure and then import it back, the arc changes back to its original appearance.

  7. Boundaries of textboxes drawn in MarvinSketch should be respected by MarvinJS, so that the text wraps properly and doesn't overlap with a structure to the right.


Higher priority feature requests (in my priority order):



  1. Add a color property to reaction arrows and electron-flow arrows, just as rectangles have already. We use colors to highlight student errors, and we refer to the colored objects in the feedback, and the fact that the feedback is referring to nonexistent colored objects can be very confusing and frustrating for students. 

  2. Allow electron-flow arrows to point to shortcut groups, especially if the electron-flow arrow starts at a bond involving that shortcut group -- e.g., the C–OMe bond in O=C–OMe.

  3. Add the capability to create a multicenter S group without adding a bond from the group to a new C atom, and allow the user to draw an electron-flow arrow from the group's attachment point to an atom

  4. Create a MarvinView equivalent: Editing not permitted, but can rotate, add explicit H atoms, clean 2D and 3D, show stereochemical descriptors.


Lower priority feature requests (no particular order):



  1. Allow the local developer to customize the list of shortcut groups.

  2. If I choose a stereobond from the Bond button group, then click on an atom button, and then I draw the atom and an associated bond by click-holding on an existing atom and dragging, the new bond should be the type that I have chosen from the Bond button group.

  3. In the various dialog boxes (View settings, Atom properties, Bond properties), enable the OK button even if I haven't changed a value. 

  4. Recognize molecule properties (of any name) in an imported MRV, and add setProperty() and getProperty() methods to the API. 

  5. Add an option to hide valence number (mrvValence) flags.

  6. Add line breaks to exported MRVs.

  7. When the implicit H display is set to heteroterm, and a C atom bears a charge, then implicit H atoms should be displayed for that C atom, like MarvinSketch does.

  8. When I select a structure or atoms, then command-click-drag to place a copy of it on the canvas, the original structure/atoms should no longer be selected, and the new copy should be selected.

  9. The head of an electron-flow arrow should not overlap with (and partly obscure) the atom symbol of the atom to which it is pointing. (See attached figure.)

  10. Make the arrow heads on electron-flow arrows larger and more visible; it is sometimes hard to tell a one-electron arrow from a two-electron arrow because the arrow head is so small. (See attached figure.)

User 870ab5b546

20-09-2016 15:05:47

New (minor) feature request: In linear cumulenes, always show the C atom. See the screen shot. 

User 870ab5b546

04-11-2016 19:46:59

Thought I'd check in again. I will be teaching a large class starting in January, and I would greatly appreciate it if you would finally take care of some of these fixes and requests, especially the following.



  1. If I replace an atom in a 3D structure with a shortcut group, the z-coordinate of the original atom is not transferred to the shortcut group. (This bug has caused students grief on exams.)

  2. Fix placement of charges so they don't overlap with bonds. (Again, this bug has caused students grief on exams.)

  3. Hypervalent heavy atoms, for example, a nine-electron Br atom, should not trigger a valence error.

  4. Allow electron-flow arrows to point to shortcut groups, especially if the electron-flow arrow starts at a bond involving that shortcut group -- e.g., the C–OMe bond in O=C–OMe.


-- Bob

ChemAxon f052bdfe3c

09-11-2016 16:35:26

I apologise for the late answer, we needed to discuss these points with the team.


1. It is a bug in our side, we will fix it.


2. The charge has a fixed position, however we create a patch around it, so it is indicate that it is not a part of the bond. Would it help if we display the carbon label, even its visibility is set off if it has charge?


3. Hypervalent heavy atoms with nine electrons are quite rare, and currently we do not have capacity to fix this issue. However, you can set the valance check off, and that case we do not indicate the error. You can even check whether it is a Br on the canvas and set the valance check off only on these cases. I hope this solution can work.


4. Since it will not lead to confusion of the electron flow starts from the bond and points to the abbreviated group which belongs to the same bond, we can add this.


Best regards,


Efi

User 870ab5b546

09-11-2016 16:55:44










ehoffmann wrote:

I apologise for the late answer, we needed to discuss these points with the team.


1. It is a bug in our side, we will fix it.


2. The charge has a fixed position, however we create a patch around it, so it is indicate that it is not a part of the bond. Would it help if we display the carbon label, even its visibility is set off if it has charge?


3. Hypervalent heavy atoms with nine electrons are quite rare, and currently we do not have capacity to fix this issue. However, you can set the valance check off, and that case we do not indicate the error. You can even check whether it is a Br on the canvas and set the valance check off only on these cases. I hope this solution can work.


4. Since it will not lead to confusion of the electron flow starts from the bond and points to the abbreviated group which belongs to the same bond, we can add this.


 



Thank you so much for your response. 


(1) Köszönöm!!!


(2) Yes, it would help if you display the C label (and the associated H atoms as well) when a C atom bears a charge. This is what MarvinSketch does (see the Figure), and MarvinSketch's behavior was the result of extensive discussion way back in the early days.


(3) Are you saying that I can show the valence check for particular atoms in a structure? For example, suppose a student draws H4C-Br(•–). Can I set Marvin JS to show the valence error for the C atom, but not the Br atom? Or are you saying that I can use JavaScript to check every atom in the drawing for a valence error, and turn off the valence error visibility flag if I find that an atom with a valence error is a hypervalent heavy atom?


(4) Köszönöm!!!


Please let me know when the promised changes are implemented.

ChemAxon f052bdfe3c

16-11-2016 13:33:37










bobgr wrote:

(3) Are you saying that I can show the valence check for particular atoms in a structure? For example, suppose a student draws H4C-Br(•–). Can I set Marvin JS to show the valence error for the C atom, but not the Br atom? Or are you saying that I can use JavaScript to check every atom in the drawing for a valence error, and turn off the valence error visibility flag if I find that an atom with a valence error is a hypervalent heavy atom?


 



Unfortunately, you cannot set the valance error visibility by atoms. However, you might can do it using structure checker webservice (see this example: https://marvinjs-demo.chemaxon.com/latest/examples/checker/example-checker.html   ) and not displaying the patch if it is on the Br atom. Could it work for you? 


Best regards,


Efi

User 870ab5b546

30-11-2016 19:49:22










ehoffmann wrote:










bobgr wrote:

(3) Are you saying that I can show the valence check for particular atoms in a structure? For example, suppose a student draws H4C-Br(•–). Can I set Marvin JS to show the valence error for the C atom, but not the Br atom? Or are you saying that I can use JavaScript to check every atom in the drawing for a valence error, and turn off the valence error visibility flag if I find that an atom with a valence error is a hypervalent heavy atom?


 



Unfortunately, you cannot set the valance error visibility by atoms. However, you might can do it using structure checker webservice (see this example: https://marvinjs-demo.chemaxon.com/latest/examples/checker/example-checker.html   ) and not displaying the patch if it is on the Br atom. Could it work for you? 


 



I looked at the example code. It seems like an awfully intricate solution for a fairly small problem. And I would rather avoid calling Web services every time a student makes a change to a structure. 


I understand that this bug is not high priority right now, but I would very much appreciate it if you would at least put it in your list of bugs to fix. Maybe you will work on the valence checker for another reason and you will be able to solve this problem at the same time. 

ChemAxon f052bdfe3c

06-12-2016 15:19:44

Could you send some scientific references/paper which can help us to identify/understand better the cases when the nine-electron Br atom is relevant. 


Thank you,


Efi

User 870ab5b546

06-12-2016 15:28:47

I don't think I can search for examples of it because of the limitations of the search engines. But I can tell you that one can encounter 9-electron Br in electron-transfer reactions to alkyl bromides (see Figure). Of course, the extra electron is really in an antibonding orital, but we don't have a way of drawing antibonding electrons, and Br can expand its octet, so we place the electron on Br and give it a formal negative charge. 


The electron transfer step in the SRN1 mechanism for substitution is another place where one might encounter 9-electron Br. 

User 870ab5b546

12-01-2017 15:19:33










bobgr wrote:

Thought I'd check in again. I will be teaching a large class starting in January, and I would greatly appreciate it if you would finally take care of some of these fixes and requests, especially the following.



  1. If I replace an atom in a 3D structure with a shortcut group, the z-coordinate of the original atom is not transferred to the shortcut group. (This bug has caused students grief on exams.)

  2. Fix placement of charges so they don't overlap with bonds. (Again, this bug has caused students grief on exams.)

  3. Hypervalent heavy atoms, for example, a nine-electron Br atom, should not trigger a valence error.

  4. Allow electron-flow arrows to point to shortcut groups, especially if the electron-flow arrow starts at a bond involving that shortcut group -- e.g., the C–OMe bond in O=C–OMe.


-- Bob












ehoffmann wrote:

I apologise for the late answer, we needed to discuss these points with the team.


1. It is a bug in our side, we will fix it.


4. Since it will not lead to confusion of the electron flow starts from the bond and points to the abbreviated group which belongs to the same bond, we can add this.


Best regards,


Efi




Here we are, another year gone by, and I am still waiting for these issues to be addressed. Any progress?

ChemAxon f052bdfe3c

18-01-2017 05:24:43

We are working on issue number 1 and quite close to the release it, however we found other cases when the third coordinate become zero, and would like to fix all of it before we start anything else. 


Best regards,


Efi

ChemAxon f052bdfe3c

02-02-2017 15:57:46

I am happy to inform you that we have released the fix on the first issue, you can download the new version from our website.


Best regards,


Efi

User 870ab5b546

02-02-2017 18:01:02

Confirmed! Thank you!


Any ETA on allowing electron-flow arrows to end at shortcut groups? And any progress on avoiding overlap of charge with bonds?