Hello,
I'm a second year CS and EE at Bristol University, England.
I have exams until the 17th of June. I know officially students are meant to start on the 23rd of May, and I would be willing to make a start then, but since I have exams around that period, I wouldn't be able to work on it full time until the 17th. Would this be a problem?
Okay, I'm thinking of putting forward 2 proposals, both from the idea's page. But before I do, I need to clarify a few things.
*Zoom Search* (and more AJAX) I'm not sure if anyone has ever looked into it, but jqPlot looks like it'd be ideal for this. It already has a zoom feature, offers a bit of "shininess", and would be quite straight forward to implement, I feel. Since pma is already implementing jQuery, using a plugin of it makes sense. jqPlot also offers a simple line of best fit. Would a few more simple mathematical functions (K-means, standard deviation, Cov?) be worth it? While I don't expect anyone to want to do anything serious in a browser, it might be worth it to have a simple way of viewing trends without having to export the database to something like matlab. Even with the extra mathematical functions, I feel like it probably wouldn't take much time to implement - i.e less than 4 weeks - and adding more AJAX seems to be the idea that naturally fits.
If this seems like a good idea I'll look more into the AJAXify part. (and probably try 1 or 2 for a contribution) Failing that, dealing with the "Browse-mode improvements", I feel that working full time on that, I could probably get through most of that list in about 4 weeks, especially since jQuery can be used. The main question here is would phpmyadmin save thing's like the last view table or SQL command in a flat file, use the database or have a choice?
The jqPlot website has a lot of examples on it ( http://www.jqplot.com/tests/ ) I'm currently working on the concept, though its still a WIP and right now only displays random data. Still, I think it should shows that doing what is required should be quite straightforward. http://mirey.pcriot.com/zoom.php
*phpMyAdmin OOP* This seems like a massive job, as it'd be rewritting the whole of PMA. Would there be any opposition to using a framework such as codeigniter? I guess the main question is, would it use none of the old code until it is finished, or would it there be a period where it would be OOP and the old procedural stuff? An argument against CI is that people would have to learn how to use CI. But I'd argue that having an arbitrary way of doing it would be more time consuming to learn, and people are likely to know how to use CI, or something similar anyway. Would it be worth splitting this down into a number of student projects, each focusing on a particular part?
I'd ultimately like to do this, though before I think about it further, I'd want to know whether I could use a framework, or if it would need to be from scratch.
Thank you, Mike
Hi Michael,
Thanks for your interest in phpMyAdmin GSoC.
2011/3/31 Michael Bundy michaeltbundy@gmail.com:
Hello,
I'm a second year CS and EE at Bristol University, England.
I have exams until the 17th of June. I know officially students are meant to start on the 23rd of May, and I would be willing to make a start then, but since I have exams around that period, I wouldn't be able to work on it full time until the 17th. Would this be a problem?
So that's about 3 weeks. It depends on how much you can do before, after and during your examens. Catching up 3 weeks of full-time (40 hours/week) development work, might be hard, but I guess you can best decide if you will be able to do it.
Okay, I'm thinking of putting forward 2 proposals, both from the idea's page. But before I do, I need to clarify a few things.
Zoom Search (and more AJAX) I'm not sure if anyone has ever looked into it, but jqPlot looks like it'd be ideal for this. It already has a zoom feature, offers a bit of "shininess", and would be quite straight forward to implement, I feel. Since pma is already implementing jQuery, using a plugin of it makes sense. jqPlot also offers a simple line of best fit. Would a few more simple mathematical functions (K-means, standard deviation, Cov?) be worth it? While I don't expect anyone to want to do anything serious in a browser, it might be worth it to have a simple way of viewing trends without having to export the database to something like matlab. Even with the extra mathematical functions, I feel like it probably wouldn't take much time to implement - i.e less than 4 weeks - and adding more AJAX seems to be the idea that naturally fits.
If this seems like a good idea I'll look more into the AJAXify part. (and probably try 1 or 2 for a contribution) Failing that, dealing with the "Browse-mode improvements", I feel that working full time on that, I could probably get through most of that list in about 4 weeks, especially since jQuery can be used. The main question here is would phpmyadmin save thing's like the last view table or SQL command in a flat file, use the database or have a choice?
The jqPlot website has a lot of examples on it ( http://www.jqplot.com/tests/ ) I'm currently working on the concept, though its still a WIP and right now only displays random data. Still, I think it should shows that doing what is required should be quite straightforward. http://mirey.pcriot.com/zoom.php
phpMyAdmin OOP This seems like a massive job, as it'd be rewritting the whole of PMA. Would there be any opposition to using a framework such as codeigniter? I guess the main question is, would it use none of the old code until it is finished, or would it there be a period where it would be OOP and the old procedural stuff?
Turning phpMyAdmin into OOP is huge, so that's why it's not in the scope of one GSoC project to rewrite everything. The idea is to rewrite small parts of the existing code, and gradually evolving to a more OOP phpMyAdmin.
No frameworks, we will just stick to OOP and design patterns.
An argument against CI is that people would have to learn how to use CI. But I'd argue that having an arbitrary way of doing it would be more time consuming to learn, and people are likely to know how to use CI, or something similar anyway. Would it be worth splitting this down into a number of student projects, each focusing on a particular part?
Basically there is one OOP project during this GSoC, so we will choose the best proposal.
I'd ultimately like to do this, though before I think about it further, I'd want to know whether I could use a framework, or if it would need to be from scratch.
Thank you, Mike
Good luck with GSoC and with submitting your proposal. Please keep in mind that you will need to submit a patch solving a bug or a feature request, as is mentioned in the 2nd point of the GSoC Applicant Guide : http://wiki.phpmyadmin.net/pma/GSoC_2011_Applicant_Guide
Kind regards,
Dieter
Hey,
On Thu, Mar 31, 2011 at 6:57 PM, Dieter Adriaenssens < dieter.adriaenssens@gmail.com> wrote:
Hi Michael,
Thanks for your interest in phpMyAdmin GSoC.
2011/3/31 Michael Bundy michaeltbundy@gmail.com:
Hello,
I'm a second year CS and EE at Bristol University, England.
I have exams until the 17th of June. I know officially students are meant
to
start on the 23rd of May, and I would be willing to make a start then,
but
since I have exams around that period, I wouldn't be able to work on it
full
time until the 17th. Would this be a problem?
So that's about 3 weeks. It depends on how much you can do before, after and during your examens. Catching up 3 weeks of full-time (40 hours/week) development work, might be hard, but I guess you can best decide if you will be able to do it.
Yeah. I could probably fit about 8-12 days of work in, so it shouldn't be too bad. My exams are quite spaced out this year which is quite nice, just means I have a long exam period.
phpMyAdmin OOP This seems like a massive job, as it'd be rewritting the whole of PMA.
Would
there be any opposition to using a framework such as codeigniter? I guess the main question is, would it use none of the old code until it
is
finished, or would it there be a period where it would be OOP and the old procedural stuff?
Turning phpMyAdmin into OOP is huge, so that's why it's not in the scope of one GSoC project to rewrite everything. The idea is to rewrite small parts of the existing code, and gradually evolving to a more OOP phpMyAdmin.
No frameworks, we will just stick to OOP and design patterns.
Ahh, okay. I understand. Makes sense. I just wanted to double check.
An argument against CI is that people would have to learn how to use CI.
But
I'd argue that having an arbitrary way of doing it would be more time consuming to learn, and people are likely to know how to use CI, or something similar anyway. Would it be worth splitting this down into a number of student projects, each focusing on a particular part?
Basically there is one OOP project during this GSoC, so we will choose the best proposal.
I'd ultimately like to do this, though before I think about it further,
I'd
want to know whether I could use a framework, or if it would need to be from scratch.
Thank you, Mike
Good luck with GSoC and with submitting your proposal. Please keep in mind that you will need to submit a patch solving a bug or a feature request, as is mentioned in the 2nd point of the GSoC Applicant Guide : http://wiki.phpmyadmin.net/pma/GSoC_2011_Applicant_Guide
Yeah. I saw that. I'm going to do some from tomorrow. I'll try a do a AJAX related feature request, and another OOP related one. I guess I should've helped out in the past =\
Kind regards,
Dieter
Thank you, Mike
2011/3/31 Michael Bundy michaeltbundy@gmail.com:
Zoom Search (and more AJAX) I'm not sure if anyone has ever looked into it, but jqPlot looks like it'd be ideal for this. It already has a zoom feature, offers a bit of "shininess", and would be quite straight forward to implement, I feel. Since pma is already implementing jQuery, using a plugin of it makes sense. jqPlot also offers a simple line of best fit. Would a few more simple mathematical functions (K-means, standard deviation, Cov?) be worth it? While I don't expect anyone to want to do anything serious in a browser, it might be worth it to have a simple way of viewing trends without having to export the database to something like matlab.
It's very unlikely you will be able to do anything useful with clustering that way, unless you extend it to more than two variables and generate group descriptions. Analysis in two dimensions can be done by simply looking at charted variables, automatic analysis is useful if you want multivariate analysis which can't be done easily that way.
Personally, I would steer away from any advanced statistical analysis - just concentrate on solid visualization of data.
Hi all,
I've just submitted a draft proposal to Google for the "Support for Procedures" project and I'm looking forward to receiving some feedback regarding this. And as I will be expanding my application during the coming days, I would be especially thankful for indications of which part of the proposal need expansion most. Many thanks in advance.
Rouslan