Category: Announcement

We recently discovered that the download counter was not correctly triggering on downloads for most of November. What that means is the majority of the downloads were not counted in November and for some possibly longer. We apologize for the error and have since corrected the problem. The download counter is now tracking again, however, we have started caching the individual SVG pages so new downloads might not appear on the individual pages unless you are logged into the site.

Have a great day!

PermaClipart would like to give a HUGE thank you to the Arweave Boost program for their generous donation to archive the clipart collection!

This post was supposed to happen back in May when we actually got the donation, but between work projects and other events the post to thank Arweave just did not happen. Thankfully what did happen was items were already being uploaded to the Arweave blockchain to be permanently stored. In fact roughly 10% of the collection has already been permanently stored on Arweave and is already available. You can follow the archive progress by viewing the transactions on the PermaClipart wallet. It has been a slow process to sumbit items to Arweave, but we are hopeful in the next little while to start taking advantage of Arweave Bundles to help speed up the archiving and keep the data and clipart SVG connected.

Again a BIG thank you to Arweave Boost for helping us archive the clipart collection and make it available for everyone for years to come!

I realize it has been awhile since I have written any updates here, but I have been busy behind the scenes making small tweaks here and there. Nothing huge and nothing that was really worth writing about. This next change though is pretty huge and I think it requires a post to announce and some of the rational behind it.

The first part of the change is a change in image formats displayed on the website. Currently all the images are PNG formats, and while that is ok, it has been a challenge. Some of the clipart in the collection has not created perfect PNG images. Render problems are the most common issue I end up trying to fix. Not to mention that real time generation of PNG images from SVG images can be pretty resource intensive and having to keep copies of PNG images for display just doubles the size of the site files overall. Not to mention that according to Can I Use basic SVG support is available for all modern browser (and no IE11 is not a modern browser IMHO).

With SVG support available in all modern browsers all clipart images displayed on the site will be the original SVG and not PNG’s generated from the SVG. The time has come to step into the modern browser era and use SVG’s. This will have some consequences since some SVG’s from the collection are HUGE but it will allow the site to not need two versions of each clipart (one SVG and one PNG) and it will mean the site will no longer have to generate PNG’s.

Second, the site offers PNG downloads of the clipart in three different sizes. Currently how the site does this, is it either generates the PNG on the fly from the SVG or the site looks to see if it has the largest version of the PNG and resizes it if needed. Depending on the SVG this can work well or in some cases not so well. Remember I mentioned that some SVG’s are HUGE? Ya those SVG’s can often cause problems and some SVG’s just don’t render out properly using the tools we currently have. Plus keeping an SVG and a PNG version of every item of clipart just doubles the size of the files needed for the website.

To fix this problem, and it is a problem, all PNG downloads will no longer be generated by us! They will be generated in client browsers using Javascript and the package called saveSvgAsPng. I tested several packages and this one seemed to do the best job and was the easiest to integrate with the website. If you download a PNG image your browser will download the SVG and render a PNG on your computer and save it. If there are problems with the render you will need to download the original SVG and create your own PNG using a tool of your choice (recommend Inkscape or GIMP but that is up to you to decide).

Third, and this is related to number two, since we are removing server side rendering of SVG’s as PNG’s, the PNG url’s will be removed from the API when this happens. If you are using the API to fetch PNG’s you will need to adjust your calls to grab the SVG’s and do the conversion on your side.

There is not a firm deadline when these changes will be implemented. I am currently testing all of them and hope to have them ready for the site by early next week. Removing the PNG url’s from the API will be the last to occur and most likely will not be until later in April.

There is also an even LARGER announcement that I will be making about the site, probably next week, that will be taking the site into the next version of the Web. More details will be available soon!

Just a quick update to mention that things are running a little behind getting API updates uploaded. The PATCH/PUT API routes are now available but I got sidetracked over the break with family and have not been able to complete the POST API route. I hope to get to it soon. Will post more once I have added the new route.

BTW: Yes I did see the latest from Openclipart, and their advancements in getting the site back up and running. I have many questions but doubt I will ever get answers, they do not seem to be very forthcoming with answers about anything and just seem to expect blind trust. It would be nice and go a long way in my opinion if they released the data on the clipart (tags, dates, etc) since that is copyrighted by them and not covered under the CC0. But I am not a lawyer so if someone knows more than me please inform me.

More to come!

I am happy to announce that nine days after stage one was complete all of the files in stage two have now been imported into the site. To explain a little more about what is included in each stage. Stage one included all the files that I was able to extract meta data from the SVG files. That usually included things like a title, date, description, tags and a creator. Not all stage one files had each meta data field but often more than one. When the title was not available in stage one files the title was created from the filename if it was possible.

Stage two files included all files that had no meta data included in the SVG file. In this case the filename occasionally had the creator, date and a title was able to be created from the filename. It is not exact since the file name scheme is inconsistent over time. Dates were harder, since there were no dates in the meta data the dates are often estimated based on the last file that had a date. When no date was able to be estimated based on the last date available by ID a default date of Jan 1, 1970 should have been assigned. Some creator names we included in the filenames, and when I noticed the name, or was able to pick it out those creator names were removed from the title and added as the creator. I am sure I missed many.

Stage three files, are all the files from stage one and stage two that have incomplete meta data and little to no clue in the filename about what they are. These files need to be done by hand. I might see if I can cross reference them with what is available in the Wayback machine but these will take longer to process. These files will also include files that have already been added that I simply missed or are having problems processing as SVG’s or images. I will probably leave many of these for a bit since I think I need a break from looking at spreadsheets of files.

Next on the site will be some clean up of things that will just make it easier to maintain the site. Here are some things on the list the next little while:

• Ability to submit changes to existing listings from the listing pages themselves when you are logged in.
• Improvement to how the quick search bar at the top and the search results page.
• Getting back to the JSON API to include the ability to POST new clipart to the site via the API and accept edits to existing items.
• Some automated processes to flag when images are not processing correctly.
• Friendlier login, registration and profile pages. This includes nicer edit pages for creators to edit listings.

If any of the original Openclipart creators would like to join it is possible for you to have the Openclipart submissions linked to your account here. This allows you to do several things:

• Edit the titles, description, dates and tags of the already included Openclipart items here on the site.
• Eventually export a CSV file of just your submissions.
• Get some statistics of downloads.
• Other suggestions? Leave a comment.

To get started you need to create an account and drop me a note with your original Openclipart user name and we will sort out what items are yours. This is still a work in progress and I am still working out some of the edit screens.

I think it is important that the original creators get credit for their work here. While that is not required under the CC0 I think the creators behind the artwork should know that their work is appreciated.

I have also updated the CSV exports of the clipart listings. The CSV file export tool will be getting a little bit of an update in the next little bit since it currently only exports the Openclipart sourced files. Now that new files can be uploaded here I need to make sure those other listings get included.

I am happy to announce that stage one of restoring Openclipart files is complete. That does not mean all the work is done, but all of the files that were scanned and had meta data embedded in them are now uploaded on the website. There are a few exceptions for files that turned out to be bad for some reason or other. The files are still available in the archive on Google Drive and at some point I hope to actually take a look at them and see if they can be fixed/restored.

Since this is a bit of a milestone I also decided to release a backup CSV of all the data from the SVG files. All the CSV backup data files will be available on Google Drive (Note: This data has been removed to save on space and $$$.). Here is what you will find:

Title,Description,Date,Creator,Tags,OCALid,Filename
75 Stroke And Fill,,2019-10-26 20:18,,Fill|Stroke,300280,id-300280---75-stroke-and-fill.svg
38 Stroke And Fill,,2019-10-26 20:18,,Fill|Stroke,300287,id-300287---38-stroke-and-fill.svg
AUDIO I2S DAC GY PCM5102,,2019-10-26 20:18,,AUDIOI|SDACGYPCM51,299722,id-299722---audio_i2s_dac_gy_pcm5102.svg
7bbn Oyv Bo W,,2019-10-26 20:18,,,300706,id-300706---7bbnoyvbow.svg
TFT LCD Screen,,2019-10-26 20:18,,145TFTLCDSCREEN,299721,id-299721---1.45_tft_lcd_screen.svg
2.1 Stroke And Fill,,2019-10-26 20:18,,Fill|Stroke,300281,id-300281---2.1-stroke-and-fill.svg
Pointing Hand,,2019-10-26 20:18,Unknown,Body|Hand|Pointing,298683,id-298683---1500150517_v2.svg

If you open the CVS file in a spreadsheet you should be able to sort and filter the data. The filenames all correspond with the folders and filenames in the Google Drive backup of the original files. Note: This data has been removed to save on space and $$$.

Some things to consider, just because the meta data in the SVG files was there does not mean it was accurate or complete. Many files would have meta data for one field but not have complete data about the file. This is why you see some very strange titles, dates that make no sense (obviously none of these files were published in October 2019). There are also several files that had similar meta data embedded in the files but it seemed to be copied from one file to the next. From what I can tell many of the newer remixes have that issue. My guess is creators copied an original SVG file, edited it and saved it but did not edit the meta data from the original file.

If you are interested in the CSV backup file feel free to download it. It is also possible for me to mass update those listing from a similar formatted CSV file, if you are a creator and know what that data should be, or just know what some are feel free to create a copy edit the items you know and send me the link via the Contact form where I can download it. I can then process the CSV file and update the listings. Just make sure the new CSV file has either the OCALid field or Filename field included.

It would be nice if the other Openclipart restoration projects would be release their data as well, but I won’t hold my breath on that happening.

Hoping to have uploads and editing on the site available by this coming weekend. Testing them now and have a few bugs left to fix.

In other news Openclipart actually tweeted for the first time since August 12th. I was shocked.

3600 followers. Path towards 4000 is open!

— Openclipart (@openclipart) November 6, 2019

The conversation from that tweet is more interesting. Particularly what does this mean?

Still in protection mode. All code working and defenses up.

— Openclipart (@openclipart) November 6, 2019

Personally I don’t think Openclipart can be trusted any longer but that is a topic for another post.